Geometry and Local Recovery of Global Minima of Two-layer Neural Networks at Overparameterization

2024-07-09 13:13:28

时间:2024年7月11日(星期四)14:30-16:00

地点:西湖大学云谷校区E4-233


主持人:理论betway必威棋牌网址 院 魏传豪

主讲人:上海交通大学 罗涛

报告主题:Geometry and Local Recovery of Global Minima of Two-layer Neural Networks at Overparameterization

报告摘要:In this talk, we investigate the geometry of the loss landscape for two-layer neural networks in the vicinity of global minima. Utilizing novel techniques, we demonstrate: (i) how global minima with zero generalization error become geometrically separated from other global minima as the sample size grows; and (ii) the local convergence properties and rate of gradient flow dynamics. Our results indicate that two-layer neural networks can be locally recovered in the regime of overparameterization.

Baidu
map