Efficient and Reliable Overlay Networks for Decentralized Federated Learning

TL;DR

This paper introduces overlay networks based on expander graphs to enhance decentralized federated learning by accelerating convergence, improving generalization, and increasing robustness, supported by theoretical analysis and empirical validation.

Contribution

It presents a novel overlay network design using spectral graph theory for DFL, with an efficient algorithm to maintain network topology amid client failures.

Findings

Accelerates convergence of DFL

Improves generalization performance

Enhances robustness to client failures

Abstract

We propose near-optimal overlay networks based on $d$-regular expander graphs to accelerate decentralized federated learning (DFL) and improve its generalization. In DFL a massive number of clients are connected by an overlay network, and they solve machine learning problems collaboratively without sharing raw data. Our overlay network design integrates spectral graph theory and the theoretical convergence and generalization bounds for DFL. As such, our proposed overlay networks accelerate convergence, improve generalization, and enhance robustness to clients failures in DFL with theoretical guarantees. Also, we present an efficient algorithm to convert a given graph to a practical overlay network and maintaining the network topology after potential client failures. We numerically verify the advantages of DFL with our proposed networks on various benchmark tasks, ranging from image classification to language modeling using hundreds of clients.