Welcome to Hexo! This is your very first post. Check documentation for more info. If you get any problems when using Hexo, you can find the answer in troubleshooting or you can ask me on GitHub.
-1 | $ hexo new "My New Post" |
More info: Writing
-1 | $ hexo server |
More info: Server
-1 | $ hexo generate |
More info: Generating
-1 | $ hexo deploy |
More info: Deployment
-
-
-
-
-
-
-
- Lixiang Chen (陈立翔 in Chinese)
- -Ph.D. Candidate, 2026
-East China Normal University
-
-
-
-
-
-
-
- RuiHao Chen (陈睿皓 in Chinese) is a Second-year Master student at School of Data Science and Engineering, East China Normal University (ECNU). He obtained his B.S. degree (2017) from College of Sciences, Shanghai University. His current interests focus on OLTP systems.
- -Master, 2024
-East China Normal University
-
-
-
-
-
-
-
- Yu Chen (陈煜 in Chinese)
- -Master, 2025
-East China Normal University
-
-
-
-
-
-
-
- Master, 2026
-East China Normal University
-
-
-
-
-
-
-
- PAN Zhicheng (潘志诚 in Chinese) is a First-year Ph.D student at School of Data Science and Engineering, East China Normal University (ECNU). He is currently an ECNU-PingCAP joint Ph.D, used to be a Data Dev intern in Taobao and a Research intern in Alibaba Cloud. He obtained both of his M.S. degree (2022) and B.S. degree (2019) from School of Computer Science and Technology, Soochow University. His current research interests focus on AI4DB.
- -Ph.D. Candidate, 2026
-East China Normal University
-
-
-
-
-
-
-
- Master, 2026
-East China Normal University
-
-
-
-
-
-
-
- Master, 2026
-East China Normal University
-
-
-
-
-
-
-
- Haoyu Wang (王浩宇)
- -Master, 2024
-East China Normal University
-
-
-
-
-
-
-
- Xin Xiong (熊鑫 in Chinese)
- -Master, 2024
-East China Normal University
-
-
-
-
-
-
-
- Master, 2026
-East China Normal University
-
-
-
-
-
-
-
- Chengcheng Yang (杨程程 in Chinese) is a research professor at School of Data Science and Engineering, East China Normal University (ECNU). Prior to joining ECNU, he was a postdoctoral researcher at School of Computer, Electrical and Mathematical Science and Engineering (CEMSE), King Abdullah University of Science and Technology. He was also a research associate at Inception Institute of Artificial Intelligence (Abu Dhabi, UAE), and a senior engineer at Shanghai Huawei Technology Co. Ltd. He obtained both of his Ph.D. degree (2017) and B.S. degree (2012) from School of Computer Science and Technology, University of Science and Technology of China. His research interests focus on database systems, storage engines, in-database machine learning, and implementing database internals with AI.
- -PhD in Computer Science, 2017
-University of Science and Technology of China
-
-
-
-
-
-
-
- Guangshun Yang (杨光舜)
- -Master, 2025
-East China Normal University
-
-
-
-
-
-
-
- Jie Yang (杨杰 in Chinese)
- -Master, 2025
-East China Normal University
-The DaSE 955 group conducts research at the intersection of Machine Learning and Databases, including AI for LSM Tree (ICDE 2023).
- - - - -
- Image credit: Unsplash
- Lorem ipsum dolor sit amet, consectetur adipiscing elit. Duis posuere tellusac convallis placerat. Proin tincidunt magna sed ex sollicitudin condimentum. Sed ac faucibus dolor, scelerisque sollicitudin nisi. Cras purus urna, suscipit quis sapien eu, pulvinar tempor diam.
- - -Slides can be added in a few ways:
-slides parameter in the front matter of the talk filestatic/ and link using url_slides parameter in the front matter of the talk fileFurther event details, including page elements such as image galleries, can be added to the body of this page.
- -
-
-
- My research interests include database systems, storage engines, in-database machine learning, and implementing database internals with AI.
- - -
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 
-
-
- 
-
-
- 
- Image credit: Unsplash
- The log-structured merge tree (LSM-tree) has been widely adopted as a backbone of modern key-value stores. However, the multiple exponentially increased levels of LSM-tree makes it suffer from high write amplification. Existing studies often improve the write performance by sacrificing the read performance, which is inefficient to make trade-offs between the update and search efficiency. In this paper, we exploit non-volatile memory (NVM) to address the write amplification issue for systems with NVM-SSD hybrid storage, and further propose a reinforcement learning method to navigate between update and search efficiency on the varying workloads. Specifically, we first propose a lightweight hot data identification method to efficiently capture access recency as well as frequency in NVM with relative large capacity. On this basis, we can eliminate different versions of frequently updated data in high-performance NVM without pushing them to SSD. To improve the data access locality and facilitate fine-grained index tuning in each level, we devise a virtual-split method to partition the key space gradually without extra write amplification. Finally, we propose a cost based Q- learning algorithm to adaptively tune the data organizations of each partition according to the changing access patterns. Ex- perimental results show that our approach outperforms existing methods by up to 2.67×.
- - - - -Supplementary notes can be added here, including code and math.
-- - - - - - - - - - - - - - - - PDF - - - - - - Cite - - - - - - - - - - - - - - - - - - - -
- - - -- - - - - - - - - - - - - - - - PDF - - - - - - Cite - - - - - - - - - - - - - - - - - - - -
- - - -
- Image credit: Unsplash
- The log-structured merge tree (LSM-tree) has been widely adopted as a backbone of modern key-value stores. However, the multiple exponentially increased levels of LSM-tree makes it suffer from high write amplification. Existing studies often improve the write performance by sacrificing the read performance, which is inefficient to make trade-offs between the update and search efficiency. In this paper, we exploit non-volatile memory (NVM) to address the write amplification issue for systems with NVM-SSD hybrid storage, and further propose a reinforcement learning method to navigate between update and search efficiency on the varying workloads. Specifically, we first propose a lightweight hot data identification method to efficiently capture access recency as well as frequency in NVM with relative large capacity. On this basis, we can eliminate different versions of frequently updated data in high-performance NVM without pushing them to SSD. To improve the data access locality and facilitate fine-grained index tuning in each level, we devise a virtual-split method to partition the key space gradually without extra write amplification. Finally, we propose a cost based Q- learning algorithm to adaptively tune the data organizations of each partition according to the changing access patterns. Ex- perimental results show that our approach outperforms existing methods by up to 2.67×.
- - - - -Supplementary notes can be added here, including code and math.
-
-
-
-
-
-
-