Bandit-based Communication-Efficient Client Selection Strategies for Federated Learning

TL;DR

This paper introduces a bandit-based client selection method for federated learning that enhances convergence speed and fairness while reducing communication costs, addressing limitations of previous biased selection strategies.

Contribution

The paper proposes UCB-CS, a novel bandit-based client selection strategy that improves convergence and fairness in federated learning with minimal communication overhead.

Findings

UCB-CS achieves faster convergence than uniform sampling.

The method reduces communication costs compared to existing biased strategies.

Client selection can be optimized for fairness in federated learning.

Abstract

Due to communication constraints and intermittent client availability in federated learning, only a subset of clients can participate in each training round. While most prior works assume uniform and unbiased client selection, recent work on biased client selection has shown that selecting clients with higher local losses can improve error convergence speed. However, previously proposed biased selection strategies either require additional communication cost for evaluating the exact local loss or utilize stale local loss, which can even make the model diverge. In this paper, we present a bandit-based communication-efficient client selection strategy UCB-CS that achieves faster convergence with lower communication overhead. We also demonstrate how client selection can be used to improve fairness.