FedClean: A Defense Mechanism against Parameter Poisoning Attacks in Federated Learning | IEEE Conference Publication | IEEE Xplore