Tackling Dynamics in Federated Incremental Learning with Variational Embedding Rehearsal

Tae Jin Park, Kenichi Kumatani, Dimitrios Dimitriadis

Federated Learning is a fast growing area of ML where the training datasets are extremely distributed, all while dynamically changing over time. Models need to be trained on client…