Key-book Project Overview
The Key-book project was initiated to serve as a valuable companion to the theoretical work "Introduction to Machine Learning Theory" (referred to as "the Guide"), authored by prominent educators such as Zhou Zhihua, Wang Wei, Gao Wei, and Zhang Lijun. The Guide fills a crucial gap in Chinese academic literature by providing an accessible introduction to machine learning theory. Its primary focus is on seven core concepts and theoretical tools crucial to understanding machine learning theory: learnability, hypothesis space complexity, generalization boundaries, stability, consistency, convergence rates, and regret bounds. While the Guide is a highly theoretical text, heavily laden with mathematical theorems and proofs, it presents certain challenges to readers without a robust mathematical background. Additionally, due to its concise nature, the Guide does not always include examples in every chapter, which can make understanding the material difficult.
To address these challenges, the contributors have developed the Key-book as a supplementary resource. The Key-book aims to provide annotations and explanations to aid readers in navigating the complexities of the Guide, thereby improving their learning experience and serving as a comprehensive record of the learning process.
Key-book Contributions
The Key-book project offers three main areas of supplemental work:
-
Concept Explanation: It introduces and explains concepts that are mentioned but not expounded upon within the Guide.
-
Proof Supplements: It elaborates on the thought processes behind some proofs, supplementing parts of the proof process that may be omitted in the Guide.
-
Case Sharing: It enriches the Guide by adding explanatory cases to facilitate a better understanding.
Since the first chapter of the Guide covers fundamental knowledge and is easily understandable, the Key-book’s content starts from the second chapter onward.
Access and Resources
- Online Reading (Real-time Updates): The Key-book is continuously updated online and can be accessed at this link.
- GitHub Repository: The project's code and resources are available on GitHub.
- Latest PDF Version: Readers can download the most recent PDF version from the releases section on GitHub.
Table of Contents
- Preface
- Chapter 1 Preliminary Knowledge
- Chapter 2 Learnability
- Chapter 3 Complexity
- Chapter 4 Generalization Boundaries
- Chapter 5 Stability
- Chapter 6 Consistency
- Chapter 7 Convergence Rates
- Chapter 8 Regret Bounds
- References
Editorial Committee
The editorial team includes:
- Chief Editors: @HaoZHAN, @zhimin-z
- Editorial Members: @leafy-lee, @MaolinWANG, @Youngfish42, @Sm1les
- Acknowledgments: @Drizzle-Zhang
Community Engagement
Interested readers can engage with the Key-book community by scanning the QR code provided to join the WeChat group or via the QQ group with the code 704768061. The respective links are available on the Key-book's online page.
Related Work
"Machine Learning" by Professor Zhou Zhihua, commonly referred to as the "Watermelon Book," is another foundational text in machine learning. Datawhale's open-source organization has further elaborated on it through the "Pumpkin Book" project. The Pumpkin Book can be read online via this link.
Licensing
The Key-book is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License, allowing others to adapt and share the material non-commercially, as long as they attribute the creators and share alike.