Efficient Detection of Byzantine Attacks in Federated Learning Using Last Layer Biases | Publicación