Abstract

We study the problem of finding a near-stationary point for smooth minimaxoptimization. The recent proposed extra anchored gradient (EAG) methods achievethe optimal convergence rate for the convex-concave minimax problem indeterministic setting. However, the direct extension of EAG to stochasticoptimization is not efficient. In this paper, we design a novel stochasticalgorithm called Recursive Anchored IteratioN (RAIN). We show that the RAINachieves near-optimal stochastic first-order oracle complexity for stochasticminimax optimization in both convex-concave andstrongly-convex-strongly-concave cases.

 

Quick Read (beta)