FedRTS: Federated Robust Pruning via Combinatorial Thompson Sampling

Federated Learning (FL) enables collaborative model training across distributed clients without data sharing, but its high computational and communication demands strain resource-constrained devices.