BriefGPT.xyz
May, 2019
深度矩阵分解中的隐式正则化
Implicit Regularization in Deep Matrix Factorization
HTML
PDF
Sanjeev Arora, Nadav Cohen, Wei Hu, Yuping Luo
TL;DR
本篇论文探讨了深度矩阵分解在矩阵补全和传感中的梯度下降隐式正则化对低秩解的影响,并发现添加深度会增强对低秩解的倾向,结果表明标准正则化的数学符号语言可能不足以完全涵盖梯度下降隐式正则化的机制。
Abstract
Efforts to understand the generalization mystery in
deep learning
have led to the belief that gradient-based optimization induces a form of
implicit regularization
, a bias towards models of low "complexity." We s
→