clients_weight = [] for c in candidates: client_weight = torch.tensor([]) diff = c.local_train() # 计算每一层参数的差值 for name, params in server.global_model.state_dict().items(): client_weight = torch.cat((client_weight, params.reshape(-1).cpu())) clients_weight.append(client...
The decentralized and privacy-preserving nature of federated learning (FL) makes it vulnerable to backdoor attacks aiming to manipulate the behavior of the... S Huang,Y Li,C Chen,... - 《Arxiv》 被引量: 0发表: 2023年 Defending Against Backdoors in Federated Learning with Robust Learning Ra...