Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Discrete Mathematic #302

Merged
merged 1 commit into from
Jan 8, 2025
Merged

Discrete Mathematic #302

merged 1 commit into from
Jan 8, 2025

Conversation

aokimi0
Copy link
Collaborator

@aokimi0 aokimi0 commented Oct 23, 2024

用AI整合了所有同学的评价,规范了文章结构,提供了更全面的信息。

@Emanual20
Copy link
Member

The unified synthetic article seems really tidy.

While to be honest, the synthetic articles (especially on "Discrete Math") summarized by LLMs lost the diversity of comments by different users, while the regular format seems "ordered". In recent study, the LLMs also face the collapse issues when they pretrained on synthetic data generated by the former version models, due to the inconsistent data distributions.

It should be mentioned that, the comments are the core of our websites. Users can find treasures from comments with different semantic style, you can even imagine what their face look when they wrote down that comments.

One solution I will try is that the unified synthetic articles contatenated with the original comments reparaphrased by LLMs.

I suggest this PR should be reviewed by more reviewers.

@aokimi0
Copy link
Collaborator Author

aokimi0 commented Oct 27, 2024

The unified synthetic article seems really tidy.

While to be honest, the synthetic articles (especially on "Discrete Math") summarized by LLMs lost the diversity of comments by different users, while the regular format seems "ordered". In recent study, the LLMs also face the collapse issues when they pretrained on synthetic data generated by the former version models, due to the inconsistent data distributions.

It should be mentioned that, the comments are the core of our websites. Users can find treasures from comments with different semantic style, you can even imagine what their face look when they wrote down that comments.

One solution I will try is that the unified synthetic articles contatenated with the original comments reparaphrased by LLMs.

I suggest this PR should be reviewed by more reviewers.

The unified synthetic article seems really tidy.

While to be honest, the synthetic articles (especially on "Discrete Math") summarized by LLMs lost the diversity of comments by different users, while the regular format seems "ordered". In recent study, the LLMs also face the collapse issues when they pretrained on synthetic data generated by the former version models, due to the inconsistent data distributions.

It should be mentioned that, the comments are the core of our websites. Users can find treasures from comments with different semantic style, you can even imagine what their face look when they wrote down that comments.

One solution I will try is that the unified synthetic articles contatenated with the original comments reparaphrased by LLMs.

I suggest this PR should be reviewed by more reviewers.

我也感觉到由LLM生成的综述文风过于简洁一致而丧失了学生评论的多样性,这可能有悖于我们网站以学生评论为核心的理念。

关于你提到的解决方案我认为是可行的,我认为可以通过恰当的提示使LLM在保持简洁性的同时尽量引用学生评论的原文。

以下是一个实例:

promt:

使用多位学生对离散数学这门课程的评论生成一篇综述性的文章,格式要求:
## 知识点与课程内容
## 上课体验
### 任明明
#### 授课风格
#### 考试相关
#### 整体评价
### 邵秀丽
#### 授课风格
#### 考试相关
#### 整体评价
要求尽量引用学生评论的原话并标注引用来源,在学生评论的多样性和综述的简洁性之间寻求平衡。

input:

(多位学生评论)

output:

知识点与课程内容

离散数学涵盖了多种基础且重要的数学知识点,包括命题逻辑、一阶谓词逻辑、集合论、笛卡尔积与二元关系、代数系统、图论等。这些知识点对于计算机科学、信息科学等学科的后续课程具有重要的基础意义。有学生指出:“前几章不算太难,但都非常重要,对于后续的课程比如人工智能导论来说非常有帮助;最难的是代数系统和图论,可能需要花点功夫。”(@natsunoshion, 2022Fall)

课程内容虽然“离散”,但学生在复习后会发现不同知识点之间存在隐晦的联系:“板块之间看上去没有任何关系,其实复习完之后总会隐约觉得毫无关联的知识点背后,是有一些只可意会不可言传的联系的。”(@FeliceRivarez, 2023Fall)

上课体验

任明明

授课风格

任明明老师的授课风格以严谨和细致著称。有学生表示:“任老师非常重视基本概念的讲解,对雨课堂习题的讲解也非常透彻。”(@natsunoshion, 2022Fall) 他善于通过清晰的推导和细致的解释来帮助学生理解复杂的数学概念。一位学生补充道:“老师讲课讲的挺好的,相比邵老师的课程安排和讲解可能都更合理一些。”(@U0Ghost, 2022Fall)

然而,一些学生对任老师的评分标准提出了意见:“老师的给分稍微差点意思,我填空判断全对,大题也证明得七七八八,但得分稀碎。”(@mingxuZhang2, 2021Fall)

考试相关

任明明老师的考试主要考察基础知识和推理能力。考前他会提供一套模拟卷,帮助学生了解考试重点:“期末考前老师会发一套模拟卷,大致能知道会考哪些题型。”(@tttran67, 2021Fall)不过,部分学生认为评分略显严格:“错了一个判断,大题证的差不多,同等完成度在隔壁可能能高2~3分。”(@U0Ghost, 2022Fall)

整体评价

总体上,任明明老师被认为是一位“认真负责”的教师。他对每一个知识点的解释非常重视,适合那些希望扎实掌握基础的学生。然而,学生普遍认为他的给分标准较为严格。尽管如此,离散数学作为计算机科学的重要基础课,其涵盖的内容对于未来的学习至关重要。

邵秀丽

授课风格

邵秀丽老师的课程风格则截然不同,被学生描述为“活泼生动”。某位学生分享:“邵老师讲课非常有激情,喜欢用生动的例子来解释复杂的概念。”(@Infrarelisxr, 2023Fall)这种互动性和生动的授课方式让课堂气氛较为轻松。然而,她的课程内容较为抽象,尤其是到了后期的群论部分,部分学生感觉难以理解:“上课是基本听不懂的。”(@FeliceRivarez, 2023Fall)

此外,邵老师会点人回答问题,但不进行点名和签到:“气温降下来之后,七十人只来二十人是常态。”(@FeliceRivarez, 2023Fall)

考试相关

邵老师的考试形式多样,考题灵活且给分宽松。一个学生提到:“邵老师给分一直不错,本人考前突击拿了90整。”(@FeliceRivarez, 2023Fall)另一位学生补充了考试题型:“平时题型较复杂,但期末难度不大,比平时作业要低。”(@Infrarelisxr, 2023Fall)

她的考试更注重理解和应用,部分学生认为考试题目“是送分题,复习的时候知识不留空白就没问题。”(@FeliceRivarez, 2023Fall)

整体评价

在学生的整体评价中,邵秀丽老师通常被认为是一位“给分慷慨”的教师。尽管课堂内容抽象且节奏较快,但她对学生的互动和课堂活跃度给予了极大的重视。她的教学更适合那些逻辑基础较好且喜欢互动式学习的学生。值得注意的是,虽然课堂内容较为分散,但考试难度相对较低,适合考前突击。

总结

离散数学这门课程因其知识点繁多且抽象,对不同类型的学生有不同的挑战。任明明老师的授课风格更为严谨,适合希望扎实掌握基础的学生;邵秀丽老师则以互动和活泼著称,更适合那些逻辑基础好、喜欢挑战的学生。总体而言,离散数学需要学生在课堂内外多花时间理解和复习,而不同教师的风格差异则为学生提供了灵活选择的空间。

这样生成的文章也许既简洁又能保持某种多样性。

@aokimi0
Copy link
Collaborator Author

aokimi0 commented Oct 27, 2024

而且,我们可以把LLM生成的综述放在正文部分,而把同学们的评论保留在评论区:)

@aokimi0 aokimi0 requested a review from Rainefly October 30, 2024 06:43
@aokimi0 aokimi0 merged commit 37b2e42 into NKUCS-ICU:main Jan 8, 2025
@aokimi0 aokimi0 deleted the dm branch January 8, 2025 11:46
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants