Title: Variables Selection of Stochastic Nonlinear System with Dynamics and Sparsity

Speaker: Wenxiao ZHAO, Academy of Mathematics and Systems Science of the Chinese Academy of Sciences

Abstract: People always hope to find objective laws from the surface and key features behind complex phenomena, therefore, the selection and extraction of key variables and factors have become important parts. How to fit the characteristics of the focus objects of system control, and incorporate the sparseness, dynamics, and nonlinearity of the system and data into this problem. We aim at the representative nonlinear ARX system (linear ARX system, Hammerstein system are the special cases) to study the above problems: first consider the non-parametric identification of such systems, construct a recursive local least squares algorithm and give a convergence analysis. Based on this, we give a criterion function for the variables selection problem of such systems and prove the estimation are both "parameter convergence" and "set convergence", and related theoretical methods have been applied to practical engineering problems.

Biography: Wenxiao Zhao graduated from the School of Mathematics, Shandong University in 2003, and received his Ph.D. from Academy of Mathematics and Systems Science of the Chinese Academy of Sciences in 2008. He then did post-doctoral research at Tsinghua University (2008-2010). He is currently an associate researcher in Academy of Mathematics and Systems Science of the Chinese Academy of Sciences.His research focuses on modeling and estimation of stochastic systems, including variables selection and sparse parameter identification, recursive estimation and adaptive control of stochastic systems, distributed estimation and optimization, and so on. Zhao Wenxiao currently serves on the editorial board of <Journal of Systems Science and Mathematical Sciences>, the editorial board of IEEE Control Systems Letters, the editorial board of IEEE CSS Conference Editorial Board, and the secretary-general of IEEE CSS Beijing.

Title: High-dimensional limit theorems for eigenvalue distributions of generalized Wishart processes

Speaker: Jian SONG, Shandong University

Abstract: In this talk, we study the first order and second order limit theorems for empirical measure of eigenvalue processes for general Wishart processes.

Biography: PhD from University of Kansas. Worked in Rutgers University at New Brunswick and Hong Kong University. National high-level talents program winner.

Research Field: Stochastic Analysis, Stochastic Partial Differential Equations, Fractional Brownian Motion, Malliavin Calculus, Random Matrix.

(Editor in Charge: Lei Hao)