BriefGPT.xyz
May, 2016
矩阵填充无虚假局部极小值
Matrix Completion has No Spurious Local Minimum
HTML
PDF
Rong Ge, Jason D. Lee, Tengyu Ma
TL;DR
本文探讨了解决基本的机器学习问题——矩阵补全的方法,并证明了普遍使用的非凸优化问题在正半定矩阵补全中没有局部极小值,可用于任意初始化的优化算法,并可在多项式时间内证明其正确性。
Abstract
matrix completion
is a basic machine learning problem that has wide applications, especially in
collaborative filtering
and
recommender systems
→