{"id":"arxiv/2609.00528","url":"https://arxiv.org/abs/2609.00528","tags":["cs.LG","math-ph","physics.chem-ph","physics.comp-ph"],"score":0,"title":"Why Multi-Layer Message Passing Works: Completeness Theory for Graph Neural Network Interatomic Potentials","domain":"cs.LG","source":"arxiv","authors":["Pingbing Ming","Han Wang"],"reviews":0,"subject":"cs.LG","summary":"We prove that the Hypergraph Neural Network, an invariant architecture with 3-body message passing, is a universal approximator for potential energy surfaces. Our main contribution is a multi-layer completeness theory. We show that $L$ layers of message passing on sparse, cutoff-based graphs achieve the same representational power as having access to the full $L$-hop neighborhood, provided the configurations are generic, satisfy an overlap condition and a connectivity condition. This provides the first rigorous justification for the common practice of using multi-layer message passing with a…","category":"research","comments":0,"disputed":false,"division":"cs","mentions":0,"discussed":0,"published":"2026-09-01","reactions":{}}