Deep understanding of machine learning: from principle to algorithmic PDF

: Network Disk Download

This book covers the rigorous theory and practical methods in machine learning, discusses the computational complexity of learning, convexity and stability, pac-Bayesian method, compression boundary and other concepts, and introduces some important algorithm paradigms, including random gradient descent, neuron network and structured output.
The book is comprehensive and thorough, suitable for high-grade undergraduate and postgraduate students with a certain foundation, and is also suitable as a reference reading for professionals and researchers who work in the IT industry for data analysis and mining.

David's monograph, "Understanding Machine Learning:from Theory to Algorithms", is a milestone in the field of machinery learning.
In recent years, machine learning is one of the most active branches in the field of artificial intelligence research, and has become an important method to solve practical problems in the field of information science, which has been applied in various fields of artificial intelligence. Machine learning is also a multidisciplinary cross-cutting field involving mathematics, automation, computer science, applied psychology, biology, and neurophysiology. The virtuous interaction brought by this interdisciplinary integration undoubtedly promotes the development and prosperity of various disciplines, including machine learning.
The content of this book is very rich, the author has not had the breadth and depth, introduced the current machine learning important theory and key algorithms. The book does not fall into the "popular science" style of the writing of materials, because the author is the authoritative experts in the field, so in the introduction of various theories and algorithms, at all times do not forget the different theories, algorithms and the author's own research results to impart to the reader, so that readers are not so rich theory and algorithm confused. In addition, it is particularly noteworthy that the first part of the book is very distinctive and a very important part. This part explores many theoretical foundations of machine learning from a higher point of view and a deeper level, and introduces the probability approximation correct (Probably approximately Correct,pac) learning theory, which is critical to guiding theoretical research and practical application. The theory is aimed at answering the question of how high the credibility and generalization of the results obtained by machine learning can be, and in a sense, only by understanding the part, is it possible to thoroughly understand and better use the content of other chapters. There is very little information about PAC learning in China, the team members encountered great difficulties in the process of translation, and our artificial intelligence and machine learning team have demonstrated and held several symposia.
This book is intended for graduate students in the fields of artificial intelligence, machine learning, pattern recognition, data mining, computer applications, bioinformatics, mathematics and statistics, and related fields of scientific and technical personnel. The aim of translation and publication is to provide a comprehensive, systematic and authoritative textbook and reference book for scholars and postgraduates who are engaged in relevant research in China. If we can do this, the translator will be very happy.
It must be stated that the translation of this book is the result of the collective efforts of the AI and machine learning research team of the Institute of Automation, CAS, members of the team Yang Xie, Kaukiuming, winsomely, Xue Wei, Wilson, Li Si, Zhang Yeheng, Zeng Fanxia, Uting, Wang Xin, Li Tao, Yang Yehui, Hu Wenrui, Zhangzhizhong, Tangyongqian, Chen Dongjie, He Zewen, Zhang Yinghua, Li Wu, Akaka, and other involved in the translation of the book, Li Sixian teacher participated in the revision and revision. Thanks to the great assistance of the mechanical industry press, the Chinese version of this book is difficult to meet with you so quickly without their enthusiastic support. In addition, the translation of the book by the National Natural Science Foundation of the Project and the project (61472423, U1135005, 61432008, 61532006, 61305018, 61402481, etc.) of the funding, hereby thank you.
In the process of translation, we strive to accurately reflect the original content, while preserving the original style. However, due to the limited level of translators, there are inevitably some irregularities in the book, and readers are urged to criticize.
Finally, I would like to dedicate the Chinese translation of this book to my doctoral tutor Wang Jue researcher! Wang Jue was very concerned about the theory, algorithm and application of machine learning, and had a unique and profound understanding of PAC learning Theory, and he inspired and led our research team to study the theory and algorithm of machine learning so that we could benefit for life.
Institute of Automation, CAS
Zhang Wensheng
April 2016 in Beijing
: Disk download Shwartz and Professor Shai, University of Waterloo, Canada Shai Shalev, associate professor at the Hebrew University of Israel

Deep understanding of machine learning: from principle to algorithmic PDF