Federated learning (FL) is a popular framework for training a model in a distributed setting, where the data is spread over a network of devices. In traditional FL, the devices communicate with a central server to share ...