Tag
This paper introduces Krum-Proxy, a selection-aware backdoor attack that bypasses distance-based robust aggregation methods like Krum in federated learning by optimizing adversarial updates to mimic benign geometry, achieving high attack success while preserving clean accuracy.
This paper reveals a non-monotonic effect of privacy on generalization error in Byzantine-robust distributed learning: in high-noise (strong privacy) regimes, increasing privacy reduces generalization error, while in low-noise (weaker privacy) regimes, increasing privacy degrades generalization.
This paper introduces 'dictator clients'—a novel class of malicious participants in federated learning capable of erasing other clients' contributions while preserving their own—and provides theoretical analysis of their impact on model convergence, including scenarios with multiple adversarial clients.