Ringleader ASGD: The First Asynchronous SGD with Optimal Time Complexity under Data Heterogeneity

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Maranjyan, Artavazd, Richtárik, Peter
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866912912490102784
author Maranjyan, Artavazd
Richtárik, Peter
author_facet Maranjyan, Artavazd
Richtárik, Peter
contents Asynchronous stochastic gradient methods are central to scalable distributed optimization, particularly when devices differ in computational capabilities. Such settings arise naturally in federated learning, where training takes place on smartphones and other heterogeneous edge devices. In addition to varying computation speeds, these devices often hold data from different distributions. However, existing asynchronous SGD methods struggle in such heterogeneous settings and face two key limitations. First, many rely on unrealistic assumptions of similarity across workers' data distributions. Second, methods that relax this assumption still fail to achieve theoretically optimal performance under heterogeneous computation times. We introduce Ringleader ASGD, the first asynchronous SGD algorithm that attains the theoretical lower bounds for parallel first-order stochastic methods in the smooth nonconvex regime, thereby achieving optimal time complexity under data heterogeneity and without restrictive similarity assumptions. Our analysis further establishes that Ringleader ASGD remains optimal under arbitrary and even time-varying worker computation speeds, closing a fundamental gap in the theory of asynchronous optimization.
format Preprint
id arxiv_https___arxiv_org_abs_2509_22860
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Ringleader ASGD: The First Asynchronous SGD with Optimal Time Complexity under Data Heterogeneity
Maranjyan, Artavazd
Richtárik, Peter
Optimization and Control
Distributed, Parallel, and Cluster Computing
Machine Learning
Asynchronous stochastic gradient methods are central to scalable distributed optimization, particularly when devices differ in computational capabilities. Such settings arise naturally in federated learning, where training takes place on smartphones and other heterogeneous edge devices. In addition to varying computation speeds, these devices often hold data from different distributions. However, existing asynchronous SGD methods struggle in such heterogeneous settings and face two key limitations. First, many rely on unrealistic assumptions of similarity across workers' data distributions. Second, methods that relax this assumption still fail to achieve theoretically optimal performance under heterogeneous computation times. We introduce Ringleader ASGD, the first asynchronous SGD algorithm that attains the theoretical lower bounds for parallel first-order stochastic methods in the smooth nonconvex regime, thereby achieving optimal time complexity under data heterogeneity and without restrictive similarity assumptions. Our analysis further establishes that Ringleader ASGD remains optimal under arbitrary and even time-varying worker computation speeds, closing a fundamental gap in the theory of asynchronous optimization.
title Ringleader ASGD: The First Asynchronous SGD with Optimal Time Complexity under Data Heterogeneity
topic Optimization and Control
Distributed, Parallel, and Cluster Computing
Machine Learning
url https://arxiv.org/abs/2509.22860