Abstract:Stochastic compositional minimax problems are prevalent in machine learning, yet there are only limited established on the convergence of this class of problems. In this paper, we propose a formal definition of the stochastic compositional minimax problem, which involves optimizing a minimax loss with a compositional structure either in primal , dual, or both primal and dual variables. We introduce a simple yet effective algorithm, stochastically Corrected stOchastic gradient Descent Ascent (CODA), which is a descent ascent type algorithm with compositional correction steps, and establish its convergence rate in aforementioned three settings. In the presence of the compositional structure in primal, the objective function typically becomes nonconvex in primal due to function composition. Thus, we consider the nonconvex-strongly-concave and nonconvex-concave settings and show that CODA can efficiently converge to a stationary point. In the case of composition on the dual, the objective function becomes nonconcave in the dual variable, and we demonstrate convergence in the strongly-convex-nonconcave and convex-nonconcave setting. In the case of composition on both variables, the primal and dual variables may lose convexity and concavity, respectively. Therefore, we anaylze the convergence in weakly-convex-weakly-concave setting. We also give a variance reduction version algorithm, CODA+, which achieves the best known rate on nonconvex-strongly-concave and nonconvex-concave compositional minimax problem. This work initiates the theoretical study of the stochastic compositional minimax problem on various settings and may inform modern machine learning scenarios such as domain adaptation or robust model-agnostic meta-learning.

Stochastically Controlled Stochastic Gradient for the Convex and Non-convex Composition problem.

Stochastically Controlled Compositional Gradient for Composition Problems

Asynchronous Proximal Stochastic Gradient Algorithm for Composition Optimization Problems

Variance Reduced Methods for Non-convex Composition Optimization

Dualityfree Methods for Stochastic Composition Optimization

Stochastic Composition Optimization of Functions without Lipschitz Continuous Gradient

Accelerated Method for Stochastic Composition Optimization With Nonsmooth Regularization

Stochastic Compositional Minimax Optimization with Provable Convergence Guarantees

Improved Sample Complexity for Stochastic Compositional Variance Reduced Gradient

Compositional Stochastic Average Gradient for Machine Learning and Related Applications

Riemannian Stochastic Gradient Method for Nested Composition Optimization

Solving Stochastic Compositional Optimization is Nearly as Easy as Solving Stochastic Optimization

Stochastic Successive Convex Approximation for Non-Convex Constrained Stochastic Optimization

Optimal Algorithms for Stochastic Multi-Level Compositional Optimization.

Asynchronous Stochastic Composition Optimization with Variance Reduction

A Smoothing Stochastic Gradient Method for Composite Optimization

Variance-Reduced Proximal Stochastic Gradient Descent for Non-convex Composite optimization.

Stochastic smoothing accelerated gradient method for nonsmooth convex composite optimization

Stochastic Recursive Momentum Method for Non-Convex Compositional Optimization

A convergence analysis of the perturbed compositional gradient flow: averaging principle and normal deviations

Stochastic subgradient for composite optimization with functional constraints