Abstract: Federated learning allows distributed clients to train a shared machine learning model while preserving user privacy. In this framework, user devices (i.e., clients) perform local iterations ...