BriefGPT.xyz
Feb, 2021
非凸矩阵分解的噪声梯度下降收敛于平坦极小值
Noisy Gradient Descent Converges to Flat Minima for Nonconvex Matrix Factorization
HTML
PDF
Tianyi Liu, Yan Li, Song Wei, Enlu Zhou, Tuo Zhao
TL;DR
本文研究了非凸矩形矩阵分解问题，通过引入噪声来解决全局极小值的不确定性，表明噪声向特定最优解施加了影响。
Abstract
Numerous empirical evidences have corroborated the importance of
noise
in
nonconvex optimization
problems. The theory behind such empirical observations, however, is still largely unknown. This paper studies this
→