Work

Towards Secure Federated Learning: The Impact of Cost-Sensitive Learning on Adversarial Resilience