Abstract
Reconciling complex, hierarchical documents, such as enterprise contracts, can present challenges, as manual review may be inefficient and probabilistic artificial intelligence models may introduce factual inaccuracies. A neuro-symbolic system may be used to address these challenges. The system can use natural language processing to extract clauses from unstructured documents and organize them into a hierarchical graph, such as a directed acyclic graph, where nodes can represent obligations and edges can represent dependencies. A deterministic validation component may then traverse this graph to perform logical and mathematical compliance checks between parent and child clauses, for example, between a master agreement and a statement of work. This approach can provide a verifiable and auditable framework for reconciling documents, which may facilitate the enforcement of rules and constraints across a document hierarchy.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Singhal, Parnika and Singhal, Dhruv, "Neuro-Symbolic System for Deterministic Reconciliation of Hierarchical Documents", Technical Disclosure Commons, (August 07, 2026)
https://www.tdcommons.org/dpubs_series/11294