无对偶、正则化和单独凸性的SDCA

Feb, 2016

SDCA without Duality, Regularization, and Individual Convexity

Shai Shalev-Shwartz

TL;DR提出了改进型的随机对偶坐标上升方法，无需显式正则化，无需依赖对偶性，甚至对于非凸损失函数，只要期望损失函数是强凸的，就可以证明收敛率是线性的。

Abstract

stochastic dual coordinate ascent is a popular method for solving regularized loss minimization for the case of convex losses. We describe variants of SDCA that do not require explicit regularization and do not r