Selective Linearization (SLIN) For Multi-block Convex Optimization

主讲人

Yu Du

Rutgers, The State University of New Jersey

Biography:

2016-Present, Instructor, Department of Management Science and Information System Department, Rutgers, The State University of New Jersey

2012- 2016, Research Assistant and Teaching Assistant at MSIS-RUTCOR, Rutgers, The State University of New Jersey

Education:

Ph.D Candidate in Operations Research, expected May 2017

Rutgers, The State University of New Jersey, New Brunswick, NJ, USA

Princeton-Rutgers Graduate Student Cooperative Exchange Program, May 2013.

Princeton University, Princeton, NJ, USA 

M.S. in Mathematical and Quantitative Finance, May 2012

Rutgers, The State University of New Jersey, New Brunswick, NJ, USA

B.S. in Economics, May 2010

Central South University, Changsha, China

Research interests:

Nonlinear Optimization, Machine Learning, Operations Research and Financial Engineering. 

讲座简介

We consider the problem of minimizing a sum of several convex non-smooth functions. We introduce a new algorithm called the selective linearization method, which iteratively linearizes all but one of the functions and employs simple proximal steps. The algorithm is a form of multiple operator splitting in which the order of processing partial functions is not fixed, but rather determined in the course of calculations. Global convergence is proved and estimates of the convergence rate are derived. Specifically, the number of iterations needed to achieve solution accuracy ε is of order O(ln(1/ε)/ε). We also illustrate the operation of the algorithm on structured regularization problems.