AI RESEARCH

Partial Model Sharing Improves Byzantine Resilience in Federated Conformal Prediction

arXiv CS.LG

ArXi:2605.11684v1 Announce Type: new We propose a Byzantine-resilient federated conformal prediction (FCP) method that leverages partial model sharing, where only a subset of model parameters is exchanged each round. Unlike existing robust FCP approaches that primarily harden the calibration stage, our method protects both the federated