D. Man

FedGAT: a backdoor attack based on global model feedback optimized triggers in federated learning

A novel backdoor attack method, termed Federated Generative Adversarial Trigger (FedGAT), which adopts a Generative Adversarial Network (GAN) framework, and can automatically produce optimized triggers that are highly correlated with the global model’s feature space, effectively reducing the “loss” in backdoor transfer...