报告人:赵文虓研究员(中国科学院数学与系统科学研究院)
报告题目:Stochastic Approximation Algorithm: General Theory and Its Applications
报告摘要：Estimating unknown parameters based on observed data widely exists in diverse areas of both theory and application. In many cases, the parameter estimation problems can be transformed into a rooting-searching problem for an unknown function f(x_0)=0. Aiming at solving the stated problem, Robbins and Monro proposed the following recursive algorithm in 1950s
x_{k+1}=x_k+a_k*y_{k+1}, k>=0
to approximate the sought-for root x_0, where a_k is the step size and y_k is the noisy observation. This kind of algorithm is called the stochastic approximation algorithm. Until now, for stochastic approximation, there have been a large amount of applications to practical problems and research works on theoretical issues.
In this talk, we will introduce the general theory for stochastic approximation algorithm, including the convergence conditions, general convergence theorem as well as the convergence rate. Then we will introduce some applications of stochastic approximation algorithm in system identification and parameter estimation, including the recursive identification for the nonlinear system with binary-valued output observations and the distributed randomized algorithm for PageRank computation.
报告人简介：Wenxiao Zhao received the BSc degree from the Department of Mathematics, Shandong University, China in 2003 and the PhD degree from the Institute of Systems Science, AMSS, the Chinese Academy of Sciences (CAS) in 2008. After this he was a Postdoc at the Department of Automation, Tsinghua University. Then, Wenxiao Zhao joined the Institute of Systems of Sciences, CAS in 2010. He now is with the Key Laboratory of Systems and Control, CAS as an Associate Professor. His research interests are in system identification and adaptive control. He serves as the General Secretary of the IEEE Control Systems Beijing Chapter, an associate editor of the Journal of Systems Science and Mathematical Sciences, and an associate editor of IEEE CSS Control Systems Letters.
报告时间：2018年7月25日(星期三)下午3:45-4:45
报告地点：科技楼(南楼)602
