Search papers, labs, and topics across Lattice.
This paper introduces DMTT (Dynamic MURMURA with Trusted Topology), a decentralized personalized federated learning protocol designed to withstand adversarial attacks on model updates and topology information. By employing evidential deep learning to down-weight unreliable peers and utilizing a confidence-weighted local topology view, DMTT effectively mitigates the influence of Byzantine adversaries, maintaining high accuracy even under significant attack scenarios. Experimental results demonstrate that DMTT achieves robust performance, sustaining accuracy above 0.862 on UCI HAR and 0.829 on PAMAP2, outperforming existing robust aggregators under various adversarial conditions.
DMTT is the only decentralized federated learning method that maintains high accuracy against adversarial attacks while ensuring Byzantine influence is effectively minimized.
In dynamic mobile decentralized federated learning (DFL), adversaries can poison both model updates and the topology information devices use to choose collaborators. We present DMTT (Dynamic MURMURA with Trusted Topology), a decentralized personalized FL protocol built on MURMURA, which uses evidential deep learning to down-weight distribution-mismatched peers, extended here to time-varying graphs under topology-manipulation attacks. Each device maintains a confidence-weighted local topology view from link-reliability estimates, signed topology claims, witness corroboration, and a Beta-distributed source-trust model, then aggregates only over a trust-screened collaborator set using a composite score fusing model compatibility, topology trust, and link reliability. We prove the screened mixing matrices confine Byzantine influence to a bounded residual $\delta_{max}$ that vanishes under perfect screening, and implement DMTT as a coordinator-free distributed system with each client running as an independent ZeroMQ process synchronized by a shared wall-clock epoch. On UCI HAR and PAMAP2, each partitioned across 100 mobile clients with Dirichlet heterogeneity, DMTT sustains honest-node accuracy above 0.862 (UCI HAR) and 0.829 (PAMAP2) across all tested adversary fractions (10 to 80%), nearly matching no-attack accuracy at low fractions and degrading gracefully toward local-only performance at extremes; static and dynamic FedAvg collapse to chance at every fraction, and robust aggregators (Krum, BALANCE, UBAR) fail to consistently beat a local-only baseline, while DMTT is the only method that clears this bar across both datasets at all fractions, with surviving Byzantine aggregation weight empirically zero throughout, consistent with $\delta_{max}$=0. The protocol runs end-to-end on real nodes via a coordinator-free ZeroMQ backend on the Melbourne Research Cloud.
