Cohomological and Thermodynamic Guardrails for Causal World Models: A Neuro-Symbolic Architecture for Verifiable Scientific Intelligence
Abstract
Contemporary foundation models operate predominantly as stateless, dissipative autoregressive token predictors lacking intrinsic representations of physical conservation laws, global causal consistency, or formal mathematical truth. We construct a neuro-symbolic causal world model governed by de Rham cohomological guardrails and non-equilibrium thermodynamic GENERIC constraints, proving that global consistency without hallucination is equivalent to the vanishing of H^1(K; F).
Keywords & Fields
Cite this work (BibTeX)
@article{deceuster2026_90-structural-compre,
title = {Cohomological and Thermodynamic Guardrails for Causal World Models: A Neuro-Symbolic Architecture for Verifiable Scientific Intelligence},
author = {De Ceuster, Peter},
year = {2026},
month = {02},
institution = {SIG Labs},
doi = {10.5281/zenodo.24252105},
url = {https://peterdeceuster.uk/papers/92-cohomological-and-thermodynamic-guardrails-for-causal-world},
note = {Full text available at https://peterdeceuster.uk/articenter/work/struct (1).pdf}
}