An Empirical Analysis of Federated Learning Models Subject to Label-Flipping Adversarial Attack