CA-HFP: Curvature-Aware Heterogeneous Federated Pruning with Model Reconstruction

ArXi:2603.12591v1 Announce Type: cross Federated learning on heterogeneous edge devices requires personalized compression while preserving aggregation compatibility and stable convergence. We present Curvature-Aware Heterogeneous Federated Pruning (CA-HFP), a practical framework that enables each client perform structured, device-specific pruning guided by a curvature-informed significance score, and subsequently maps its compact submodel back into a common global parameter space via a lightweight reconstruction.