federated learning (FL), as an emerging collaborative learning paradigm, has
garnered significant attention due to its capacity to preserve privacy within
distributed learning systems. In these systems, clients collaboratively train a
unified neural network model using their local data