Abstract: Federated learning (FL) faces significant challenges in scenarios with nonindependent and identically distributed (non-IID) data distributions across participating clients. Traditional ...