Stochastic momentum ADMM for nonconvex and nonsmooth optimization with application to PnP algorithm

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Deng, Kangkang, Zhang, Shuchang, Wang, Boyu, Jin, Jiachen, Zhou, Juan, Wang, Hongxia
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913800837398528
author Deng, Kangkang
Zhang, Shuchang
Wang, Boyu
Jin, Jiachen
Zhou, Juan
Wang, Hongxia
author_facet Deng, Kangkang
Zhang, Shuchang
Wang, Boyu
Jin, Jiachen
Zhou, Juan
Wang, Hongxia
contents This paper proposes SMADMM, a single-loop Stochastic Momentum Alternating Direction Method of Multipliers for solving a class of nonconvex and nonsmooth composite optimization problems. SMADMM achieves the optimal oracle complexity of $\mathcal{O}(ε^{-3/2})$ in the online setting. Unlike previous stochastic ADMM algorithms that require large mini-batches or a double-loop structure, SMADMM uses only $\mathcal{O}(1)$ stochastic gradient evaluations per iteration and avoids costly restarts. To further improve practicality, we incorporate dynamic step sizes and penalty parameters, proving that SMADMM maintains its optimal complexity without the need for large initial batches. We also develop PnP-SMADMM by integrating plug-and-play priors, and establish its theoretical convergence under mild assumptions. Extensive experiments on classification, CT image reconstruction, and phase retrieval tasks demonstrate that our approach outperforms existing stochastic ADMM methods both in accuracy and efficiency, validating our theoretical results.
format Preprint
id arxiv_https___arxiv_org_abs_2504_08223
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Stochastic momentum ADMM for nonconvex and nonsmooth optimization with application to PnP algorithm
Deng, Kangkang
Zhang, Shuchang
Wang, Boyu
Jin, Jiachen
Zhou, Juan
Wang, Hongxia
Optimization and Control
Numerical Analysis
65K05, 65K10, 90C05, 90C26, 90C30
This paper proposes SMADMM, a single-loop Stochastic Momentum Alternating Direction Method of Multipliers for solving a class of nonconvex and nonsmooth composite optimization problems. SMADMM achieves the optimal oracle complexity of $\mathcal{O}(ε^{-3/2})$ in the online setting. Unlike previous stochastic ADMM algorithms that require large mini-batches or a double-loop structure, SMADMM uses only $\mathcal{O}(1)$ stochastic gradient evaluations per iteration and avoids costly restarts. To further improve practicality, we incorporate dynamic step sizes and penalty parameters, proving that SMADMM maintains its optimal complexity without the need for large initial batches. We also develop PnP-SMADMM by integrating plug-and-play priors, and establish its theoretical convergence under mild assumptions. Extensive experiments on classification, CT image reconstruction, and phase retrieval tasks demonstrate that our approach outperforms existing stochastic ADMM methods both in accuracy and efficiency, validating our theoretical results.
title Stochastic momentum ADMM for nonconvex and nonsmooth optimization with application to PnP algorithm
topic Optimization and Control
Numerical Analysis
65K05, 65K10, 90C05, 90C26, 90C30
url https://arxiv.org/abs/2504.08223