BriefGPT.xyz
Jun, 2015
自适应随机原始对偶坐标下降算法处理可分离鞍点问题
Adaptive Stochastic Primal-Dual Coordinate Descent for Separable Saddle Point Problems
HTML
PDF
Zhanxing Zhu, Amos J. Storkey
TL;DR
本文提出了一种基于随机块坐标下降和自适应步长的原始-对偶更新框架，用于解决具有可分结构的凸-凹鞍点问题，并在正则化的经验风险最小化问题上进行了实证分析，结果表明该方法具有较快的收敛速度和良好的表现。
Abstract
We consider a generic
convex-concave saddle point problem
with separable structure, a form that covers a wide-ranged machine learning applications. Under this problem structure, we follow the framework of
primal-dual up
→