Now Professor Zhang Zhihua of Peking University has published the development course and enlightenment of machine learning in the Newsletter of China Computer Federation 20 17 [1]. This paper recommends the learning methods of machine learning in detail.
Machine learning integrates technology, science and art. It is different from traditional artificial intelligence and is the core of modern artificial intelligence. It involves statistics, optimization, matrix analysis, theoretical computer, programming, distributed computing and so on. Therefore, it is suggested to strengthen the courses of probability, statistics and matrix analysis on the basis of the existing undergraduate courses of computer specialty. The following are suggestions on the specific curriculum and related teaching materials: 1. To strengthen the basic course of probability and statistics, it is suggested to adopt the fourth edition of Probability Theory and Mathematical Statistics co-authored by Morris H. DeGroot and Mark J. Schervish. Douban Link-Probability and Statistics
2. In the course of linear algebra, strengthen the content of matrix analysis. It is suggested to use Gilbert Strong's Introduction to Linear Algebra in the textbook. Gilbert Strong has been teaching linear algebra at MIT, and his online video course is a classic. Later, it was suggested to set up matrix calculation, and the textbook Numerical Linear Algebra written by Trefethen N. Lloyd and David Bau lll was adopted. 3. Set up machine learning courses. There are many classic books on machine learning, but most of them are not suitable for undergraduates. Recently, John D. Kelleher, Brian Mac Namee and others published "The Basis of Machine Learning for Predictive Data Analysis". Or the third edition of Statistical Pattern Recognition co-authored by Andrew R. Webb and Keith D. Copsey, which is more suitable for undergraduates. At the same time, it is suggested to set up practical links in the course, so that students can try to apply machine learning methods to some specific problems.