<ul><li>Federated continual learning (FCL) allows each client to continually update its knowledge from task streams.</li><li>FCL needs to address spatial data heterogeneity between clients and temporal data heterogeneity between tasks.</li><li>The proposed Federated Tail Anchor (FedTA) overcomes parameter-forgetting and output-forgetting using trainable Tail Anchor and frozen output features.</li><li>FedTA also includes Input Enhancement, Selective Input Knowledge Fusion, and Best Global Prototype Selection for improved performance in downstream tasks.</li></ul>

Addressing Spatial-Temporal Data Heterogeneity in Federated Continual Learning via Tail Anchor

Discover more