Momentum-Based Acceleration for Non-convex Stochastic Optimisation
We consider stochastic non-convex optimization problems that arise in several applications including machine learning and the stochastic gradient Hamiltonian Monte Carlo (SGHMC) algorithm to solve them. We obtain the first finite-time global convergence guarantees for SGHMC in the context of both empirical and population risk minimization. Our results show SGHMC can achieve acceleration on a class of non-convex problems compared to overdamped Langevin MCMC approaches such as the stochastic gradient Langevin dynamics. This is joint work with Mert Gurbuzbalaban and Lingjiong Zhu.
主讲人简介：
Xuefeng Gao is currently an Assistant Professor in the Department of Systems Engineering and Engineering Management at the Chinese University of Hong Kong. He joined the department in November 2013 after obtaining his Ph.D. in Operations Research from Georgia Institute of Technology. His research interests include applied probability and stochastic processes, queueing theory, algorithmic and high frequency trading.
欢迎广大师生前来参加！
