【学术报告】A framework for analyzing variance reduced stochastic gradient methods and a new one

发布者:孙毅 发布时间:2021-11-29 浏览次数:555

时间:2021-12-9 下午1430

题目:A framework for analyzing variance reduced stochastic gradient methods and a new one

报告人:梁经纬,副教授,上海交通大学自然科学研究院


摘要:

Over the past years, variance reduced stochastic gradient methods have become increasingly popular, not only in the machine learning community, but also other areas including inverse problems and mathematical imaging to name a few. However, despite the varieties of variance reduced stochastic gradient descent methods, their analysis varies from each other. In this talk, I will first present a unified framework, under which we manage to abstract different variance reduced stochastic gradient methods into one. Then I will introduce a new stochastic method for composed optimization problems, and illustrate its performance via several imaging problems.


简介:

Research interests:  

Non-smooth Optimization, Mathematical Imaging and Data Science.  

Short bio:  

Tenure-track Associate Professor, Institute of Natural Science and School of Mathematical Sciences, Shanghai Jiao Tong University.

Lecturer in Mathematical Data Science (2020.11-2021.06), School of Mathematical Sciences, Queen Mary University of London.

Postdoc research associate (2017.04-2020.08), Department of Applied Mathematics and Theoretical Physics, University of Cambridge.

Ph.D. in Mathematics (2016.10) GREYC, ENSICAEN and University of Caen Normandy.

M.S. in Mathematics (2013.03), School of Mathematics & Institute of Natural Sciences, Shanghai Jiao Tong University.

B.S. in Electrical & Information Engineering (2010.06), School of Telecommunication and Information Engineering, Nanjing University of Posts and Telecommunications.