2022/12/05 by Lesi Chen, Chen, Lesi, Haishan Ye +3 · 5 citations
Computer Science · Engineering · #Distributed Control Multi-Agent Systems #FOS: Computer and information sciences #FOS: Mathematics #Machine Learning (cs.LG) #Optimization and Control (math.OC) #Sparse and Compressive Sensing Techniques #Stochastic Gradient Optimization Techniques
paper · pdf · doi:10.48550/arxiv.2212.02387
openalex publication_date 2022/12/05 · openalex created_date 2022/12/19 · openalex updated_date 2026/07/28
This paper studies the stochastic nonconvex-strongly-concave minimax optimization over a multi-agent network. We propose an efficient algorithm, called Decentralized Recursive gradient descEnt Ascent Method (DREAM), which achieves the best-known theoretical guarantee for finding the ε-stationary points. Concretely, it requires O(min (κ3ε-3,κ2 √(N) ε-2 )) stochastic first-order oracle (SFO) calls and O(κ2 ε-2) communication rounds, where κ is the condition number and N is the total number of individual functions. Our numerical experiments also validate the superiority of DREAM over previous methods.