Introduction to Artificial Intelligence book list

Author | Wang Tianyi

SOURCE | Geek Time column "Artificial Intelligence Basic Course"

Doctor of Engineering, associate professor of AI Collection book, with the text attached to the PDF version of the link.

Machine Learning Chapter

In machine learning, the first recommendation is the work of two domestic authors: "Statistical learning methods" by Dr. Hangyuan Li and "machine learning" by Professor Zhou Zhihua.

"Statistical learning Method" adopts "total-division-Total" structure, after combing the basic concept of statistical learning, systematically and comprehensively introduces the 10 main methods of statistical learning, and finally summarizes and compares these algorithms. This book is based on mathematical formulas, the introduction of each method is given a detailed mathematical deduction, almost no nonsense, so the reader's mathematical background also put forward a higher demand.

In contrast, machine learning covers a broader range of features that are more introductory in nature and help to understand the panorama of machine learning. The book covers the basic ideas, scope of application, advantages and disadvantages of almost all the algorithms in machine learning and the main ways of realization, interspersed with a large number of easy-to-understand examples.

If the "Statistical learning method" wins in depth, then "machine learning" wins in the breadth. On the premise of breadth, we can continue to dig deep according to the rich references provided in machine learning.

After reading the above two books, we can read some classics. The classic works of Tom Mitchell, the machine learning, the Chinese version of the translation called "Computer Learning." This book was written in 1997, although it is difficult to cover the latest advances in machine learning, but the basic theory and core algorithms are still so pertinent, after all, classical theory can withstand the test of time. The focus of the book is also the breadth, not involving a large number of complex mathematical deduction, is a relatively ideal primer books. The author has said on his homepage that the book is going to be a new version, supplemented by some chapters, and may be expected to appear in the next two years.

Another classic is Elements of statistical learning, Trevor Hastie, who published the second edition in 2016. This book is not translated, only a photocopy. A master's book does not scare people with a lot of complex mathematical formulas (except for books that are designed for algorithmic deduction), which is no exception. It emphasizes the connotation and extension of various learning methods, and it is probably more important to understand their application scenarios and development direction in the context of the method than the concrete deduction.

The Pattern recognition and machine learning of the non-Christopher Bishop appearing on the finale is the genus. Published in 2007, there is no Chinese translation, perhaps the reason is that such a gaudy magnum translation out of the Midnight oil to spend how many nights. This book is characterized by the machine learning as a whole, regardless of the frequency-based approach or Bayesian method, whether it is a regression model or a classification model, is only a different aspect of the problem. The author is able to open God's perspective and incorporate a wide range of machine learning into a huge network, unfortunately, most readers can't keep up with his strategically advantageous position (including myself).

Finally recommended is David J C MacKay information theory, inference and Learning Algorithms, written in 2003, the Chinese version is called "Information theory, reasoning and learning algorithm." The author of this book is a versatile scientist, this book is not a monograph on machine learning, but a number of related disciplines into a furnace, the content of a very wide range. Compared to the front-face textbook, reading this book is like chatting with the author, he will throw all kinds of questions to make you think. A wide range of topics makes the reading experience of the book not easy, but can be an adjustment to broaden your horizons.

Math article

1. Linear algebra

Two foreign textbooks are recommended. One is Gilbert Strang Introduction to Linear Algebra, the English version in 2016 to the fifth edition, no Chinese translation. This is a conceptual interpretation of the intuitive image of the basic concept of abstraction, coupled with a large number of linear algebra in various fields of practical application, the learner is very friendly. The author has set up a video course on the OCW of MIT, and is equipped with a series of electronic resources, such as problem solving and simulated questions.

The second is David C Lay's Linear Algebra and its applications, the English version in 2015 also went to the fifth edition, the Chinese version is called "Linear algebra and its application", corresponding to the original book fourth edition. This book, through the basic concepts of vectors and linear equations, introduces the basic concepts in the line generation, emphasizes the algebraic meanings and geometrical meanings behind the formulas, and also has a large number of application examples, which is very helpful for understanding the basic concepts.

2. Probability theory

Basic reading can choose Sheldon M Ross of a first Course in probability, the English version in 2013 to the Nineth edition (18 immediately to the tenth edition), the Chinese version of the "Basic Probability theory Tutorial", corresponding to the original book Nineth version, also has a photocopy of English. This book throws away the measure, discusses the probability question from the angle of the central limit theorem, the explanation of the concept is more popular, the book also contains the massive close contact Life application example and the example question problem.

Another difficult reading is Edwin Thompson Jaynes probability theory:the Logic of Science, the book has no Chinese translation, the photocopy of the "theory of probability meditation" is also out of print. This book is the author of the posthumous, spent half a century to complete, from the name can be seen as a book of God. The author discusses the probability, Bayesian probability and statistical inference based on frequency, and introduces the subject of probability theory into the framework of mathematical logic. If you read this book, you must be prepared to burn your brain.

3. Mathematical Statistics

Basic readings can be selected by Chen Shiyin academician of the "Mathematical Statistics course". The question of whether statistics is scientific is still consensus, but its important role in machine learning is beyond doubt. Chen Lao's book focuses on the concept of statistics and ideas, trying to teach the use of statistical ideas to observe and analyze the ability of things, this is very valuable.

Advanced reading can choose Roger Casella Statistical Inference, as the author died in 2012, the second edition of 2001 became the swan song. The Chinese translation is called "Statistical inference" and there is also a photocopy. This book contains some of the content of probability theory, and introduces the basic problems of statistical inference, parameter estimation, variance regression and other statistics.

4. Optimization theory

You can refer to Stephen Boyd's convex optimization, the Chinese version of which is called "Convex optimization". Although this book is scary, but not very readable, mainly for practical applications rather than theoretical proof, many of the methods widely used in machine learning can find the source here.

5. Information theory

It is recommended that Thomas Cover and Jay A Thomas co-authored Elements of information theory,2006 to the second edition of the "Information base". This book takes into account the breadth and depth, although not a voluminous but dry, clear the information theory of the basic concepts of the physical connotation, but to smooth reading needs a certain mathematical basis. In addition, this book emphasizes the application of information theory in communication.

About the author

Wang Tianyi, Ph. D., University of Posts and Telecommunications, Associate professor, College of Big Data and information engineering, Guizhou Province, member of 3D Digital Medical Society of Guizhou. During reading, the main research direction is continuous variable quantum communication theory and system, presided over and participated in a number of national/provincial research projects, as the first author published 5 SCI papers.

The research focuses on big data and artificial intelligence, including big data applications based on the internet of things and neural networks and machine learning. In addition to the technical field, the development of artificial intelligence and future trends are also in-depth thinking, there is the "AI Revolution" book.

PDF Link

Machine Learning Chapter

1. Machine learning Http://www.cs.ubbcluj.ro/~gabis/ml/ml-books/McGrawHill%20-%20Machine%20Learning%20-Tom%20Mitchell.pdf
2. Elements of statistical learning Https://web.stanford.edu/~hastie/Papers/ESLII.pdf
3. Pattern Recognition and machine learning http://users.isr.ist.utl.pt/~wurmd/Livros/school/Bishop%20-%20Pattern% 20recognition%20and%20machine%20learning%20-%20springer%20%202006.pdf
4. Information theory, inference and learning algorithms http://www.inference.org.uk/itprnn/book.pdf

Math article

1. Introduction to Linear Algebra https://math.mit.edu/~gs/linearalgebra/linearalgebra5_Preface.pdf
2. Linear Algebra and its Applications http://www.zuj.edu.jo/download/linear-algebra-and-its-applications-david-c-lay-pdf/
3. A first Course in probability (8th edition) http://julio.staff.ipb.ac.id/files/2015/02/Ross_8th_ed_English.pdf
4. Probability theory:the Logic of science http://www.med.mcgill.ca/epidemiology/hanley/bios601/GaussianModel/ Jaynesprobabilitytheory.pdf
5. Statistical inference Https://fsalamri.files.wordpress.com/2015/02/casella_berger_statistical_inference1.pdf
6. Convex optimization Https://web.stanford.edu/~boyd/cvxbook/bv_cvxbook.pdf
7. Elements of information theory Http://www.cs-114.org/wp-content/uploads/2015/01/Elements_of_Information_Theory_ Elements.pdf

The above content, from "Artificial Intelligence basic Course". If you are interested in AI, or consider transforming the AI field. This column is exactly what you need. There are seven major modules in the column:

• Module one: Mathematical basis
• Module Two: The Main method of machine learning
• Module three: Artificial neural network
• Module four: Deep learning
• Module Five: Examples of neural networks
• Module Six: Artificial intelligence beyond deep learning
• Module VII: Application Scenarios

Introduction to Artificial Intelligence book list