This open access textbook provides a rigorously structured, method-oriented guide to mastering core machine learning methods essential for understanding and applying modern artificial intelligence (AI) technologies. In an era where AI is transforming every industry, Machine Learning Methods (Second Edition) systematically presents the most foundational and widely used techniques across four key branches: supervised learning, unsupervised learning, deep learning, and reinforcement learning.
The book is clearly organized around algorithmic methods—such as GBDT, the EM algorithm, Transformer models, diffusion models, and PPO—that have remained central to machine learning despite rapid advancements in the field. Through concise mathematical formulations, intuitive explanations, and practical examples, it offers deep insights into over 40 essential techniques. Each volume provides a focused overview, followed by chapters that explain one or two key methods, making the content accessible for both comprehensive study and targeted reference.
Designed for advanced undergraduate and graduate students, educators, and AI professionals, this textbook serves both as a learning resource and a long-term reference. It assumes foundational knowledge in calculus, linear algebra, probability, and computer science, and supports readers in developing a structured understanding of machine learning that is both theoretical and application-oriented. Whether exploring why Transformers have revolutionized natural language processing or how PPO optimizes decision-making in reinforcement learning, this book is intended to both inform and inspire further exploration.
About the Author :
Hang Li is a recognized authority in machine learning, natural language processing, and information retrieval. A Fellow of the Association for Computational Linguistics, ACM, and IEEE, he is also a Distinguished Member of the China Computer Federation. Dr. Li holds a Ph.D. in Computer Science from the University of Tokyo and began his career as a researcher at NEC Corporation. He later advanced to senior roles as a Research Manager at Microsoft Research Asia, Director of Huawei Noah's Ark Lab, Director of ByteDance AI Lab, and Head of Research at ByteDance.
He has authored more than 160 papers in premier conferences and journals, including NeurIPS, ICML, ACL, SIGIR, and JMLR, and has served on the editorial boards and program committees of many leading venues in these fields. Dr. Li is also a recipient of the CCF-ACM AI Award and several Best Paper awards, and he holds more than 60 U.S. patents related to real-world AI applications. He has contributed to the development of influential products such as Microsoft SharePoint and Jinri Toutiao.
His book, Statistical Learning Methods, is widely known as the "Blue Bible" of machine learning in China, and the present volume builds on this legacy by offering a comprehensive, method-driven perspective on core machine learning technologies. Known for his precision in theory and clarity in writing, Dr. Li brings both depth and accessibility to the study of machine learning.
