Report title: Optimization Methods for Acceleration: Adaptation, Subsampling and Subproblem-Solving
Reporter: Professor Jiang Bo Shanghai University of Finance and Economics
Reporting time: 9:00-9:40 AM on January 4, 2021
Report location: Tencent Conference ID: 320 247 940
Conference password: 9999
School contact: Li Xinxin email@example.com
In this talk we present a suite of accelerated algorithms for solving optimization models arising from applications such as machine learning. In general, popular optimization methods for solving such problems include the high-order tensor approximation approach, which requires the knowledge on some problem parameters. To make such methods practical, one will need to find ways of implementation without such knowledge. Furthermore, when the objective in such models involves a large number of terms, a subsampling strategy will be adopted. We discuss methods that exhibit an accelerated iteration bound while maintaining the traits of being adaptive and allowing subsampling.
Jiang Bo is a professor at the School of Information Management and Engineering at Shanghai University of Finance and Economics. He received his Ph.D. from the Department of Industrial and Systems Engineering at the University of Minnesota in September 2013, and his supervisor is Professor Zhang Shuzhong. The main research areas include optimization theory, portfolio optimization, signal processing, image processing, etc. He has published many papers in Operations Research, Mathematics of Operations Research, Mathematical Programming, SIAM Journal on Optimizatoin, etc. He is currently a reviewer of mathmatical reviews under the American Mathematical Society, and has served as an anonymous reviewer for well-known journals such as Management Science, Mathematics of Operations Research, SIAM Journal on Optimization, etc. Served as a summer researcher at Whitebox Advisors, a US hedge fund company (with assets under management of approximately US$4.5 billion), engaged in research on robust portfolio investment. The main recent research topic is the theory of data-driven optimization and its application in revenue management and operation management.