李宏毅深度学习教程LeeDL-Tutorial(苹果书)
李宏毅老师是台湾大学的教授,其《机器学习》(2021年春)是深度学习领域经典的中文视频之一。李老师幽默风趣的授课风格深受大家喜爱,让晦涩难懂的深度学习理论变得轻松易懂,他会通过很多动漫相关的有趣例子来讲解深度学习理论。李老师的课程内容很全面,覆盖了到深度学习必须掌握的常见理论,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,培养深度学习的直觉,对于想入门深度学习又想看中文讲解的同学是非常推荐的。
本教程主要内容源于《机器学习》(2021年春),并在其基础上进行了一定的原创。比如,为了尽可能地降低阅读门槛,笔者对这门公开课的精华内容进行选取并优化,对所涉及的公式都给出详细的推导过程,对较难理解的知识点进行了重点讲解和强化,以方便读者较为轻松地入门。此外,为了丰富内容,笔者在教程中选取了《机器学习》(2017年春) 的部分内容,并补充了不少除这门公开课之外的深度学习相关知识。
ℹ️ 李宏毅老师推荐:
纸质版
京东扫码购买 |
当当扫码购买 |
豆瓣评分:https://book.douban.com/subject/36997460/
最新版PDF下载
地址:https://github.com/datawhalechina/leedl-tutorial/releases
国内地址(推荐国内读者使用):链接: https://pan.baidu.com/s/1KUOlEMlPi5I4b5ys7aJVjw 提取码: fmuk
内容介绍
- Introduction @Wang Qi
- Deep Learning @Wang Qi
- Local Minima and Saddle Points
- Training Tips
- Adaptive Learning Rate
- Loss Functions for Classification Problems
- Normalization
- Convolutional Neural Networks and Self-Attention Mechanism @Wang Qi
- Convolutional Neural Networks
- Self-Attention Mechanism
- Recurrent Neural Networks @Wang Qi
- Transformer @Wang Qi
- Transformer
- Generative Models @Yang Yiyuan
- Basics of Generative Adversarial Networks
- Theory of Generative Adversarial Networks and Wasserstein Generative Adversarial Networks
- Evaluation of Generative Adversarial Networks and Conditional Generative Adversarial Networks
- Recurrent Generative Adversarial Networks
- Self-Supervised Learning @Wang Qi
- Sesame Street Models
- BERT
- GPT-3
- Concepts and Applications of Autoencoders @Jiang Ji
- Diffusion Models @Wang Qi
- Adversarial Attack @Yang Yiyuan
- Basic Concepts of Adversarial Attack
- White-box Attacks vs Black-box Attacks
- Passive Defense vs Active Defense
- Explainable Artificial Intelligence @Yang Yiyuan
- Concepts and Cases of Explainable Artificial Intelligence
- Local Explainability in Explainable Artificial Intelligence
- Global Explainability in Explainable Artificial Intelligence
- Transfer Learning @Wang Qi
- Domain Adaptation
- Domain Adversarial Training
- Deep Reinforcement Learning @Wang Qi
- Lifelong Learning @Jiang Ji
- Catastrophic Forgetting
- Mitigation of Catastrophic Forgetting
- Network Compression @Wang Qi
- Pruning and Lottery Hypothesis
- Knowledge Distillation
- Meta Learning @Yang Yiyuan
- Concept of Meta Learning
- Example Algorithms of Meta Learning
- Applications of Meta Learning
- ChatGPT @Yang Yiyuan
- Misunderstandings about ChatGPT
- Key Technologies Behind ChatGPT - Pre-Training
- Research Problems Introduced by ChatGPT
Supporting Code
Click or go to the Homework
folder on the webpage to access the supporting code.
Additional Resources
- For readers interested in Reinforcement Learning, you can read Mushroom Book EasyRL
- For readers interested in Visual Reinforcement Learning, you can read Awesome Visual RL
Contributors
Qi Wang PhD student at Shanghai Jiao Tong University |
Yiyuan Yang PhD student at the University of Oxford |
John Jim Master's degree from Peking University |
Citation Information
@misc{wang2023leedltutorial,
title = {李宏毅深度学习教程},
year = {2023},
author = {王琦,杨毅远,江季},
url = {https://github.com/datawhalechina/leedl-tutorial}
}
@misc{wang2023leedltutorialen,
title = {Deep Learning Tutorial by Hung-yi Lee},
year = {2023},
author = {Qi Wang,Yiyuan Yang,Ji Jiang},
url = {https://github.com/datawhalechina/leedl-tutorial}
}
If this tutorial has helped you, please support it by clicking the Star :star: button on the top right corner of the page, thanks :blush:!
If you need to reprint the content of this tutorial, please indicate the source: https://github.com/datawhalechina/leedl-tutorial.
Acknowledgements
Special thanks to @Sm1les, @LSGOMYP, FuWeiru for their help and support on this project.
Additionally, thank you everyone for your attention to the LeeDL-Tutorial.
Follow Us
Scan the QR code below to follow the WeChat official account: Datawhale, reply with the keyword "李宏毅深度学习", and you can join the "LeeDL-Tutorial Readers' Group."
LICENSE
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.