BriefGPT.xyz
Oct, 2017
质量传输正则化
Regularization via Mass Transportation
HTML
PDF
Soroosh Shafieezadeh-Abadeh, Daniel Kuhn, Peyman Mohajerin Esfahani
TL;DR
本文提出了使用分布式鲁邦优化的思想来作为正则化技术以及对现有技术提供新的概率解释。通过选择半径,可以保证最坏情况下的预期损失提供了对测试数据的上限置信度,从而提供新的泛化界限。
Abstract
The goal of regression and classification methods in
supervised learning
is to minimize the
empirical risk
, that is, the expectation of some loss function quantifying the prediction error under the empirical dist
→