Saved in:
Bibliographic Details
Main Authors: Zhu, Mingxi, Ye, Yinyu
Format: Preprint
Published: 2022
Subjects:
Online Access:https://arxiv.org/abs/2208.09735
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915288788762624
author Zhu, Mingxi
Ye, Yinyu
author_facet Zhu, Mingxi
Ye, Yinyu
contents Distributed optimization algorithms are widely used in machine learning. This paper investigates how a small amount of data sharing can improve their performance. Focusing on general linear models, we analyze the effects of data sharing on both primal and primal-dual optimization methods. Our contributions are threefold. First, from a theoretical perspective, we show that minimal data sharing improves algorithmic performance by shifting data from less favorable to more favorable structures. Contrary to the common belief that data heterogeneity is always harmful, we prove that while heterogeneity generally slows convergence in primal methods such as FedAvg and distributed PCG, it can accelerate convergence in primal-dual consensus algorithms like distributed ADMM, Fed-ADMM, and EXTRA by enriching dual dynamics. This reveals a form of duality in how heterogeneity affects different algorithm families. Second, building on this insight, we design a meta-algorithm for minimal data sharing, adaptable to both primal and primal-dual methods. We show that with as little as 1 percent shared data, convergence can be significantly accelerated across machine learning tasks. Finally, we argue from a broader perspective that even limited collaboration can yield large synergies, an idea that transcends the optimization context. Our findings provide both theoretical and practical guidance for improving distributed learning through minimal cooperation and motivate further exploration of cross-agent collaboration in solving complex global learning problems.
format Preprint
id arxiv_https___arxiv_org_abs_2208_09735
institution arXiv
publishDate 2022
record_format arxiv
spellingShingle How a Small Amount of Data Sharing Benefits Distributed Optimization and Learning : The Upside of Data Heterogeneity
Zhu, Mingxi
Ye, Yinyu
Optimization and Control
90C06 (Primary), 90C25, 68U04 (Secondary)
Distributed optimization algorithms are widely used in machine learning. This paper investigates how a small amount of data sharing can improve their performance. Focusing on general linear models, we analyze the effects of data sharing on both primal and primal-dual optimization methods. Our contributions are threefold. First, from a theoretical perspective, we show that minimal data sharing improves algorithmic performance by shifting data from less favorable to more favorable structures. Contrary to the common belief that data heterogeneity is always harmful, we prove that while heterogeneity generally slows convergence in primal methods such as FedAvg and distributed PCG, it can accelerate convergence in primal-dual consensus algorithms like distributed ADMM, Fed-ADMM, and EXTRA by enriching dual dynamics. This reveals a form of duality in how heterogeneity affects different algorithm families. Second, building on this insight, we design a meta-algorithm for minimal data sharing, adaptable to both primal and primal-dual methods. We show that with as little as 1 percent shared data, convergence can be significantly accelerated across machine learning tasks. Finally, we argue from a broader perspective that even limited collaboration can yield large synergies, an idea that transcends the optimization context. Our findings provide both theoretical and practical guidance for improving distributed learning through minimal cooperation and motivate further exploration of cross-agent collaboration in solving complex global learning problems.
title How a Small Amount of Data Sharing Benefits Distributed Optimization and Learning : The Upside of Data Heterogeneity
topic Optimization and Control
90C06 (Primary), 90C25, 68U04 (Secondary)
url https://arxiv.org/abs/2208.09735