DriftGuard: Mitigating Asynchronous Data Drift in Federated Learning

TL;DR

DriftGuard is a federated learning framework that efficiently manages asynchronous data drift by separating shared and local parameters, enabling targeted retraining and reducing costs while maintaining high accuracy.

Contribution

It introduces a Mixture-of-Experts inspired architecture for federated learning that separates global and local parameters, enabling efficient adaptation to asynchronous data drift.

Findings

Reduces retraining cost by up to 83%

Matches or exceeds state-of-the-art accuracy

Improves accuracy per unit retraining cost by up to 2.3x

Abstract

In real-world Federated Learning (FL) deployments, data distributions on devices that participate in training evolve over time. This leads to asynchronous data drift, where different devices shift at different times and toward different distributions. Mitigating such drift is challenging: frequent retraining incurs high computational cost on resource-constrained devices, while infrequent retraining degrades performance on drifting devices. We propose DriftGuard, a federated continual learning framework that efficiently adapts to asynchronous data drift. DriftGuard adopts a Mixture-of-Experts (MoE) inspired architecture that separates shared parameters, which capture globally transferable knowledge, from local parameters that adapt to group-specific distributions. This design enables two complementary retraining strategies: (i) global retraining, which updates the shared parameters when system-wide drift is identified, and (ii) group retraining, which selectively updates local parameters for clusters of devices identified via MoE gating patterns, without sharing raw data. Experiments across multiple datasets and models show that DriftGuard matches or exceeds state-of-the-art accuracy while reducing total retraining cost by up to 83%. As a result, it achieves the highest accuracy per unit retraining cost, improving over the strongest baseline by up to 2.3x. DriftGuard is available for download from https://github.com/blessonvar/DriftGuard.