1 paper
Youssef Allouah, Rachid Guerraoui, Nirupam Gupta +3
Robust federated learning aims to maintain reliable performance despite the presence of adversarial or misbehaving workers. While state-of-the-art (SOTA) robust distributed gradien…