Bayesian Ying-Yang System, Best Harmony Learning, and Five Action Circling¶
Lei Xu
Chair Professor
The Chinese University of Hong Kong
Abstract :
Proposed in 1995 and systematically developed over fifteen years, Bayesian Ying-Yang (BYY) learning is a statistical approach for an intelligent system via two complementary Bayesian representations of a joint distribution on the external observation X and its inner representation R, called BYY system. A Ying-Yang best harmony principle is proposed for learning all the unknowns in the system, in help of an implementation featured by a five action circling. BYY learning provides not only a general framework that accommodates typical learning approaches from a unified perspective but also a new road that leads to improved model selection criteria, automatic model selection during learning, and coordinated implementation of Ying based model selection and Yang based learning regularization. This talk introduces BYY learning principles, implementing techniques, and typical learning algorithms, in a comparison with other algorithms, particularly with the EM algorithm as a benchmark. These algorithms are summarized in a unified Ying-Yang alternation procedure with major parts in a same expression while differences simply characterized by few options.
Bio:
Lei Xu, chair professor of Chinese Univ Hong Kong, Chang Jiang Chair Professor of Peking Univ, IEEE Fellow (2001-) and Fellow of International Association for Pattern Recognition (2002-), and Academician of European Academy of Sciences (2002-). He completed his Ph.D thesis at Tsinghua Univ by the end of 1986, then joined Dept. Math, Peking Univ in 1987 first as a postdoc and then exceptionally promoted to associate professor in 1988 and to a full professor in 1992. During 1989-93 he worked at several universities in Finland, Canada and USA, including Harvard and MIT. He joined CUHK in 1993 as senior lecturer, as professor in 1996 and chair professor in 2002. He has published a number of well-cited papers on neural networks, statistical learning, and pattern recognition, e.g., his papers got over 3200 citations according to SCI and over 5500 citations according to Google Scholar (GS), with the first 10 papers scored over 2000 (SCI) and 3600 (GS). One single paper has scored 750 (SCI) and 1211 (GS). He served as associate editor for several journals, past governor of international neural network society (INNS), a past president of APNNA, and a member of Fellow committee of IEEE CI Society. Also, he has received several national and international academic awards (e.g., 1993 National Nature Science Award, 1995 INNS Leadership Award and 2006 APNNA Outstanding Achievement Award).