神经网络中的泛化特性识别

Sep, 2018

Identifying Generalization Properties in Neural Networks

Huan Wang, Nitish Shirish Keskar, Caiming Xiong, Richard Socher

TL;DR通过PAC-Bayes范式描述的解的局部特性，证明了模型泛化能力与Hessian、Lipschitz常数和参数的尺度有关，并提出了泛化指标及相应的算法。

Abstract

While it has not yet been proven, empirical evidence suggests that model generalization is related to local properties of the optima which can be described via the hessian. We connect →