跳到主要导航 跳到搜索 跳到主要内容

Generalization Performance of Empirical Risk Minimization on Over-Parameterized Deep ReLU Nets

  • Xi'an Jiaotong University
  • The University of Sydney

科研成果: 期刊稿件 › 文章 › 同行评审

1 引用 (Scopus)

摘要

In this paper, we study the generalization performance of global minima of empirical risk minimization (ERM) on over-parameterized deep ReLU nets. Using a novel deepening scheme for deep ReLU nets, we rigorously prove that there exist perfect global minima achieving optimal generalization error rates for numerous types of data under mild conditions. Since over-parameterization of deep ReLU nets is crucial to guarantee that the global minima of ERM can be realized by the widely used stochastic gradient descent (SGD) algorithm, our results present a potential way to fill the gap between optimization and generalization of deep learning.

源语言英语
页(从-至)1978-1993
页数16
期刊IEEE Transactions on Information Theory
卷71
期3
DOI
出版状态已出版 - 2025

学术指纹

探究 'Generalization Performance of Empirical Risk Minimization on Over-Parameterized Deep ReLU Nets' 的科研主题。它们共同构成独一无二的学术指纹。

引用此