Free Energy Manifold: Score-Based Inference for Hybrid Bayesian Networks

ArXi:2605.09839v1 Announce Type: cross A central finding is the mode-bridge artifact: standard conditional energy models can create low-energy ridges between separated modes of the same class, producing overconfident posteriors at off-data interior points. We analyze this failure and propose valley regularization, an off-data calibration term that res near-uniform posteriors in such regions while preserving in-data fit.