arxivstat.MLcs.LG2026-07-06
Decomposition for Bayesian Networks: Local and Parallel Inference
Probabilistic inference in high-dimensional Bayesian networks is difficult because exact manipulation of the joint distribution scales exponentially with network size. We propose a decomposition framework based on directed convex subgraphs and introduce a minimal d-decomposition…