Customizing Proximal Point Algorithm for Convex Programming: Applications to Sparse and Low-rank Models
主 题: Customizing Proximal Point Algorithm for Convex Programming: Applications to Sparse and Low-rank Models
报告人: Prof. Xiaoming Yuan (Hong Kong Baptist University)
时 间: 2013-12-02 14:00-15:00
地 点: 全9教室，北京国际数学研究中心全斋(主持人：文再文)
The proximal point algorithm (PPA) is a fundamental method in optimization, playing significant roles both theoretically and algorithmically. In this talk, I will show that the PPA with customized choices of its proximal parameter can be very efficient for solving some convex programming problems. Some applications to sparse and/or low-rank driven models will be shown, including the basis pursuit, imaging denoising/deblurring, and video surveillance problems.
