Title: machine learning based on cognitive axiomatic
Abstract: In the era of big data, due to demand-driven applications, a large number of new machine learning methods continue to produce. These new algorithms based on different theories, the relationship between them is extremely complex, user learning algorithms demanding. However, the learning ability of children is high, but not the current theory of machine learning to master. Whether the proposed set of machine learning theory in line with human cognition, is currently a serious problem.
The report attempts to present a unified machine learning based on cognitive axiomatic framework, the basic assumption is: return what kind of, like what kind; like what kind of, what kind of return.
The machine learning theory can deduce three design principles of classification method, in a uniform manner reinterpret the data dimensionality reduction, density estimation, regression, clustering and classification issues, and in line with the principles of cognitive daily life.
To the sword, the current Artificial Intelligence Research Institute of Beijing Jiaotong University, executive vice president, two professors, traffic data analysis and mining Beijing Key Laboratory Director, CCF Fellow, CCF Artificial Intelligence and Professional Committee Moshishibie director-designate (2020- 2023), CAAI director, deputy director of the professional Committee CAAI machine learning. Author of monographs "Machine Learning: from axioms to Algorithms", as written in CAAI organization textbook "Introduction to Artificial Intelligence" executive editor.