Federated Learning under Distributed Concept Drift
This addresses a largely unexplored challenge in federated learning for scenarios with time-varying data distributions across distributed clients, representing an incremental advance in handling data heterogeneity.
The paper tackles the problem of federated learning under distributed concept drift, where drifts occur staggered across clients, by proposing new clustering algorithms for drift adaptation, achieving accuracy comparable to an oracle with ground-truth knowledge.
Federated Learning (FL) under distributed concept drift is a largely unexplored area. Although concept drift is itself a well-studied phenomenon, it poses particular challenges for FL, because drifts arise staggered in time and space (across clients). To the best of our knowledge, this work is the first to explicitly study data heterogeneity in both dimensions. We first demonstrate that prior solutions to drift adaptation that use a single global model are ill-suited to staggered drifts, necessitating multiple-model solutions. We identify the problem of drift adaptation as a time-varying clustering problem, and we propose two new clustering algorithms for reacting to drifts based on local drift detection and hierarchical clustering. Empirical evaluation shows that our solutions achieve significantly higher accuracy than existing baselines, and are comparable to an idealized algorithm with oracle knowledge of the ground-truth clustering of clients to concepts at each time step.