Understanding Byzantine Robustness in Federated Learning with A Black-box Server

Open in new window