报告题目 (Title)：基于值函数的双层超参数选择差分凸算法
报告人 (Speaker)：张进 副教授 (南方科技大学)
报告时间：2023年09月08日（周五）15:30
参会方式：线上，腾讯会议514-871-643
邀请人：周安娃
主办部门：理学院数学系
报告摘要：
Bilevel Optimization gains significant attention recently due to its various applications. Gradient-based methods guarantee theoretical convergence to stationary solutions when the lower level of the bilevel program is strongly convex (LLSC) and smooth (LLS) for fixed upper-level variable values. In this talk, we present a sequentially convergent Value Function based Difference-of-Convex Algorithm with inexactness (VF-iDCA). We show that this algorithm achieves stationary solutions without LLSC and LLS assumptions for bilevel programs from a broad class of hyperparameter tuning applications. Extensive numerical experiments justify our theoretical results and show that the proposed VF-iDCA yields superior performance.
