Federated Learning under Arbitrary Communication Patterns

                 Dmitrii Avdyukhin 1 Shiva Prasad Kasiviswanathan 2

            Abstract                 The canonical federated learning problem involves learning
                                 a single, global statistical model from data stored on lots
   Federated Learning is a distributed learning set-      of remote devices. In particular, the goal is typically to
   ting where the goal is to train a centralized model     minimiz ...