学术报告:Markov-Chain Stochastic Optimization Methods

题目:Markov-Chain Stochastic Optimization Methods

主讲人:印卧涛 教授,加州大学洛杉矶分校

日期:2018年6月25日(星期一)

时间:上午10:00 - 11:30

地点:数据科学与计算机学院 A201

主持:凌青 教授

 

摘要:This talk introduces Markov-chain stochastic optimization methods. They are useful in reinforcement learning, dynamical system identification, and other optimization applications where Markov chains naturally arise. In addition, they can be used to design communication-efficient decentralized algorithms. We show that stochastic gradient methods and random block-coordinate methods can still converge when we sample according to a Markov chain.

 

个人介绍:Wotao Yin received his Ph.D. in operations research from Columbia University in 2006. He is currently a Professor with the Department of Mathematics, University of California, Los Angeles. His research interests include computational optimization and its applications in signal processing, machine learning, and other data science problems. He has received the NSF CAREER award in 2008, the Alfred P. Sloan Research Fellowship in 2009, and the Morningside Gold Medal in Applied Mathematics in 2016, and has coauthored five papers receiving best paper-type awards. He invented fast algorithms for sparse optimization and has been leading the research of optimization algorithms for large-scale problems.