<ul data-eligibleForWebStory="true"><li>Federated Learning (FL) faces challenges from concept drift where client data distributions change over time.</li><li>Existing FL methods focus on real drift but struggle with virtual and label drift, leading to catastrophic forgetting.</li><li>FedDAA is introduced as a dynamic clustered FL framework to address multi-source concept drift by incorporating modules for cluster number determination, real drift detection, and concept drift adaptation.</li><li>Experiments demonstrate that FedDAA outperforms state-of-the-art methods with significant accuracy improvements on datasets like Fashion-MNIST, CIFAR-10, and CIFAR-100.</li></ul>

FedDAA: Dynamic Client Clustering for Concept Drift Adaptation in Federated Learning

Discover more