When to Trust Aggregated Gradients: Addressing Negative Client...

Federated Learning has become a widely-used framework which allows learning a global model on decentralized local datasets under the condition of protecting local data privacy. However, federated...