Abstract: Federated learning (FL) algorithms usually sample a fraction of clients in each round (partial participation) when the number of participants is large and the server's communication ...
Abstract: The success of most robust model estimation methods heavily relies on their used data sampling algorithms. This letter proposes a novel sampling algorithm, called Guided Sampling by ...