National Taiwan University: Machine Learning by Hung-yi Lee
Course Overview
- University: National Taiwan University
- Prerequisites: Proficiency in Python
- Programming Language: Python
- Course Difficulty: 🌟🌟🌟🌟
- Estimated Hours：80 hours
Professor Hung-yi Lee, a professor at National Taiwan University, is known for his humorous and engaging teaching style. He often incorporates fun elements like Pokémon into his slides, making the learning experience enjoyable.
Although labeled as a machine learning course, the breadth of topics covered is impressive. The course includes a total of 15 labs covering Regression, Classification, CNN, Self-Attention, Transformer, GAN, BERT, Anomaly Detection, Explainable AI, Attack, Adaptation, RL, Compression, Life-Long Learning, and Meta Learning. This wide coverage allows students to gain insights into various domains of deep learning, helping them choose areas for further in-depth study.
Don't be overly concerned about the difficulty of the assignments. All assignments come with example code from teaching assistants, guiding students through data processing, model building, and more. Students are required to make modifications based on the provided code. This presents an excellent opportunity to learn from high-quality code, and the assignments serve as valuable resources for those looking to breeze through course projects.
Course Resources
- Course Websites：https://speech.ee.ntu.edu.tw/~hylee/ml/2022-spring.php
- Course Videos：https://speech.ee.ntu.edu.tw/~hylee/ml/2022-spring.php
- Course Textbook: N/A
- Course Assignments：https://speech.ee.ntu.edu.tw/~hylee/ml/2022-spring.php, 15 labs covering a wide range of deep learning domains
National Taiwan University: Machine Learning by Hung-yi Lee
Course Overview
- University: National Taiwan University
- Prerequisites: Proficiency in Python
- Programming Language: Python
- Course Difficulty: 🌟🌟🌟🌟
- Estimated Hours：80 hours
Professor Hung-yi Lee, a professor at National Taiwan University, is known for his humorous and engaging teaching style. He often incorporates fun elements like Pokémon into his slides, making the learning experience enjoyable.
Although labeled as a machine learning course, the breadth of topics covered is impressive. The course includes a total of 15 labs covering Regression, Classification, CNN, Self-Attention, Transformer, GAN, BERT, Anomaly Detection, Explainable AI, Attack, Adaptation, RL, Compression, Life-Long Learning, and Meta Learning. This wide coverage allows students to gain insights into various domains of deep learning, helping them choose areas for further in-depth study.
Don't be overly concerned about the difficulty of the assignments. All assignments come with example code from teaching assistants, guiding students through data processing, model building, and more. Students are required to make modifications based on the provided code. This presents an excellent opportunity to learn from high-quality code, and the assignments serve as valuable resources for those looking to breeze through course projects.
The 2025 version of the course has undergone a reform of the course content, focusing more on RAG, AI Agent, LLM all sorts of fancier content; it differs greatly from the 2023 version and previous versions
Course Resources
- Course Websites：Spring2023, Spring2025
- Course Videos：Spring2023, Spring2025
- Course Textbook: N/A
- Course Assignments：Spring2023(15 labs covering a wide range of deep learning domains), Spring2025 (focus on LLM related work like AI Agent)
国立台湾大学：李宏毅机器学习
课程简介
- 所属大学：國立台灣大學
- 先修要求：熟练掌握 Python
- 编程语言：Python
- 课程难度：🌟🌟🌟🌟
- 预计学时：80 小时
李宏毅老师是国立台湾大学的教授，其风趣幽默的授课风格深受大家喜爱，并且尤其喜欢在 PPT 中插入宝可梦等动漫元素，是个非常可爱的老师。
这门课挂着机器学习的牌子，但其课程内容之广实在令人咋舌，其作业一共包含 15 个 lab，分别是 Regression、Classification、CNN、Self-Attention、Transformer、GAN、BERT、Anomaly Detection、Explainable AI、Attack、Adaptation、 RL、Compression、Life-Long Learning 以及 Meta Learning。可谓是包罗万象，能让学生对于深度学习的绝大多数领域都有一定了解，从而可以进一步选择想要深入的方向进行学习。
大家也大可不必担心作业的难度，因为所有作业都会提供助教的示例代码，帮你完成数据处理、模型搭建等，你只需要在其基础上进行适量的修改即可。这也是一个学习别人优质代码的极好机会，大家需要水课程大作业的话，这里也是一个不错的资料来源。
课程资源
- 课程网站：https://speech.ee.ntu.edu.tw/~hylee/ml/2022-spring.php
- 课程视频：https://speech.ee.ntu.edu.tw/~hylee/ml/2022-spring.php，每节课的链接参见课程网站
- 课程教材：无
- 课程作业：https://speech.ee.ntu.edu.tw/~hylee/ml/2022-spring.php，15 个 lab，几乎覆盖了主流深度学习的所有领域
国立台湾大学：李宏毅机器学习
课程简介
- 所属大学：國立台灣大學
- 先修要求：熟练掌握 Python
- 编程语言：Python
- 课程难度：🌟🌟🌟🌟
- 预计学时：80 小时
李宏毅老师是国立台湾大学的教授，其风趣幽默的授课风格深受大家喜爱，并且尤其喜欢在 PPT 中插入宝可梦等动漫元素，是个非常可爱的老师。
这门课挂着机器学习的牌子，但其课程内容之广实在令人咋舌，其作业一共包含 15 个 lab，分别是 Regression、Classification、CNN、Self-Attention、Transformer、GAN、BERT、Anomaly Detection、Explainable AI、Attack、Adaptation、 RL、Compression、Life-Long Learning 以及 Meta Learning。可谓是包罗万象，能让学生对于深度学习的绝大多数领域都有一定了解，从而可以进一步选择想要深入的方向进行学习。
大家也大可不必担心作业的难度，因为所有作业都会提供助教的示例代码，帮你完成数据处理、模型搭建等，你只需要在其基础上进行适量的修改即可。这也是一个学习别人优质代码的极好机会，大家需要水课程大作业的话，这里也是一个不错的资料来源。
2025年版课程的课程内容发生改革，更加侧重于RAG、AI Agent、LLM种种更fasion的内容；与2023版及之前版本差异极大
课程资源
- 课程网站：Spring2023, Spring2025
- 课程视频：Spring2023, Spring2025，每节课的链接参见课程网站
- 课程教材：无
- 课程作业：Spring2023 (5 个 lab，几乎覆盖了主流深度学习的所有领域；部分作业colab上可能无法打开，这时候可以参考弘毅老师的github), Spring2025 (主要关注 AI Agent 等 LLM 相关领域)