Title: Cognition-based Machine Learning Axiomatization
Time: 10: 00am, Friday, January 10, 2020
Venue: Administration Building, Room 601
Organizer: School of Artificial Intelligence
Abstract: In the era of big data, a large number of new machine learning methods, driven by application requirements, are constantly generated. These new algorithms have different theoretical basis, and the relationship between them is extremely complicated, and place extremely high demands on the users of the algorithms. Whether we can propose a set of machine learning theories in line with human cognition is an urgent problem to be solved.
This talk attempts to propose a unified cognition-based axiomatized framework for machine learning, and its basic assumptions are “ to which category, like which category; like which category, to which category ”. This machine learning theory can deduce three design principles of classification methods, and reinterpret data dimensionality reduction, density estimation, regression, clustering and classification in a unified way, which is consistent with the cognitive principles in daily life.
Biography: Jian Yu is a professor and the Executive Deputy Dean of Institute of Artificial Intelligence, Beijing Jiaotong University. He is the Director of Beijing Key Laboratory of Traffic Data Analysis and Mining. He is the Director of the Artificial Intelligence and Pattern Recognition Professional Committee of CCF from 2020. He is the Deputy Secretary-General and Executive Director of CAAI, and Deputy Director of Machine Learning Professional Committee of CAAI. He has written a book “Machine Learning: From Axioms to Algorithms”, and is the Executive Editor of the textbook “Introduction to Artificial Intelligence” by CAAI.
