Exact Penalty Method for Federated Learning

TL;DR

This paper introduces FedEPM, an exact penalty method for federated learning that improves communication efficiency, reduces computational complexity, mitigates stragglers' effects, and enhances data privacy, with proven convergence and strong numerical results.

Contribution

It proposes FedEPM, a novel exact penalty-based algorithm specifically designed for federated learning to address key challenges and improve overall performance.

Findings

Proven convergence of FedEPM.

High numerical performance demonstrated.

Addresses communication, computation, stragglers, and privacy issues.

Abstract

Federated learning has burgeoned recently in machine learning, giving rise to a variety of research topics. Popular optimization algorithms are based on the frameworks of the (stochastic) gradient descent methods or the alternating direction method of multipliers. In this paper, we deploy an exact penalty method to deal with federated learning and propose an algorithm, FedEPM, that enables to tackle four critical issues in federated learning: communication efficiency, computational complexity, stragglers' effect, and data privacy. Moreover, it is proven to be convergent and testified to have high numerical performance.