{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,17]],"date-time":"2026-09-17T06:03:55Z","timestamp":1789625035850,"version":"build-2803163510"},"reference-count":30,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,9,17]],"date-time":"2026-09-17T00:00:00Z","timestamp":1789603200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Artif. Intell."],"abstract":"<jats:p>The ability to access specific operational guidelines in unstructured technical documentation is a crucial and unsolved problem in safety sensitive areas like aerospace, clinical care, and regulatory compliance, where approximate retrieval and hallucination are unacceptable. Current systems of Retrieval Augmented Generation (RAG) relying on dense vector embeddings are unable to offer the deterministic accuracy such environments require. In this paper, a neurosymbolic framework is provided to overcome this limitation through the translation of natural language documents into three complementary representations of symbolic knowledge: a SQLite (Lightweight SQL Database Engine) database of SAT (Boolean Satisfiability) - based propositional reasoning Conjunctive Normal Form (CNF) clauses, a directed property graph relational traversal and a geometric spatial index which represents the quantitative conditional rules as axis aligned hyper rectangles in an eight dimensional parameter space. Each of the three representations is built on the same LLM (Large Language Model) - based rule extraction pipeline at the same time during offline stage, where there is no embedding of vectors of any type. Online three path reasoning architecture combines SAT at query time entailment checking, knowledge graph traversal, and exact geometric containment testing \u2014 translating the operational state of a user to a parameter space point and recalling all rules with hyper rectangular regions containing the point. The outputs of all the three paths are combined before constrained answer generation, basing each answer on logical, relational, and numerical evidence. Because retrieval is performed using mathematically precise symbolic procedures rather than similarity approximation, the SAT, graph, and geometric reasoning components are deterministic and fully explainable; large language model calls are still used for rule extraction and answer synthesis, so the system as a whole is hybrid rather than fully deterministic, and a faithfulness filter together with context-confined generation substantially mitigate, rather than formally guarantee, hallucination. The experimental findings show that this tri-representation strategy markedly enhances retrieval accuracy and reliability compared to embedding-based baselines in complicated operational query conditions, developing a scalable base of rule-based decision support in high stakes areas.<\/jats:p>","DOI":"10.3389\/frai.2026.1922508","type":"journal-article","created":{"date-parts":[[2026,9,17]],"date-time":"2026-09-17T05:33:55Z","timestamp":1789623235000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["A hybrid RAG and rule-based reasoning framework for technical document analysis"],"prefix":"10.3389","volume":"9","author":[{"given":"Esther Rachel","family":"Thomas","sequence":"first","affiliation":[{"name":"School of Computer Science and Engineering, Vellore Institute of Technology","place":["Chennai, India"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Niranchna","family":"Natarajan","sequence":"additional","affiliation":[{"name":"School of Computer Science and Engineering, Vellore Institute of Technology","place":["Chennai, India"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Keerthana","family":"Jayaprakashan","sequence":"additional","affiliation":[{"name":"School of Computer Science and Engineering, Vellore Institute of Technology","place":["Chennai, India"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"A.","family":"Malini","sequence":"additional","affiliation":[{"name":"School of Computer Science and Engineering, Vellore Institute of Technology","place":["Chennai, India"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,9,17]]},"reference":[{"key":"ref1","doi-asserted-by":"publisher","first-page":"5611","DOI":"10.1109\/ISCC.2008.4625611","article-title":"A new approach to content-based file type detection proceedings of the IEEE international conference on information security and cryptology","volume":"2010","author":"Amirani","year":"2010","journal-title":"arXiv"},{"key":"ref2","doi-asserted-by":"publisher","first-page":"3781","DOI":"10.1016\/j.procs.2024.09.178","article-title":"A survey on retrieval-augmented generation with large language models","volume":"246","author":"Arslan","year":"2024","journal-title":"ACM Comput. Surv."},{"key":"ref3","doi-asserted-by":"publisher","first-page":"100179","DOI":"10.1016\/j.nlp.2025.100179","article-title":"A synergistic multi-stage retrieval-augmented generation architecture for boosting context relevance in data science literature proceedings of an international conference on artificial intelligence and data science","volume":"13","author":"Aytar","year":"2025","journal-title":"Nat. Lang. Proc. J."},{"key":"ref4","doi-asserted-by":"publisher","first-page":"05150","DOI":"10.48550\/arXiv.2004.05150","article-title":"Longformer: the long-document transformer","volume":"2020","author":"Beltagy","year":"2020","journal-title":"arXiv"},{"key":"ref5","doi-asserted-by":"publisher","first-page":"1870","DOI":"10.18653\/v1\/P17-1171","article-title":"Reading Wikipedia to answer open-domain questions proceedings of ACL","volume":"1","author":"Chen","year":"2017","journal-title":"arXiv"},{"key":"ref6","doi-asserted-by":"publisher","first-page":"14887","DOI":"10.48550\/arXiv.2401.14887","article-title":"The power of noise: redefining retrieval for retrieval-augmented generation systems, proceedings of EMNLP","volume":"2024","author":"Cuconasu","year":"2024","journal-title":"arXiv"},{"key":"ref7","doi-asserted-by":"publisher","first-page":"394","DOI":"10.1145\/368273.368557","article-title":"A machine program for theorem-proving","volume":"5","author":"Davis","year":"1962","journal-title":"Commun. ACM"},{"key":"ref8","author":"De Moura","year":"2008"},{"key":"ref9","doi-asserted-by":"publisher","first-page":"10997","DOI":"10.48550\/arXiv.2312.10997","article-title":"Retrieval-augmented generation for large language models: a survey","volume":"2024","author":"Gao","year":"2024","journal-title":"arXiv"},{"key":"ref10","doi-asserted-by":"publisher","first-page":"47","DOI":"10.1145\/602259.602266","article-title":"R-trees: a dynamic index structure for spatial searching proceedings of the ACM SIGMOD international conference on management of data","volume":"14","author":"Guttman","year":"1984","journal-title":"ACM SIGMOD Record"},{"key":"ref11","doi-asserted-by":"publisher","first-page":"3929","DOI":"10.48550\/arXiv.2002.08909","article-title":"REALM: retrieval-augmented language model pre-training proceedings of ICML","volume":"2020","author":"Guu","year":"2020","journal-title":"arXiv"},{"key":"ref12","doi-asserted-by":"publisher","first-page":"12981","DOI":"10.48550\/arxiv.2506.12981","article-title":"SymRAG: efficient neuro-symbolic retrieval through adaptive query routing","volume":"2025","author":"Hakim","year":"2025","journal-title":"arXiv"},{"key":"ref13","doi-asserted-by":"publisher","first-page":"494","DOI":"10.1109\/TNNLS.2021.3070843","article-title":"A survey on knowledge graphs: representation, acquisition, and applications","volume":"33","author":"Ji","year":"2022","journal-title":"IEEE Trans. Neural Networks Learn. Syst."},{"key":"ref14","doi-asserted-by":"publisher","first-page":"5983","DOI":"10.48550\/arXiv.2410.05983","article-title":"Long-context large language models meet retrieval-augmented generation: overcoming challenges for long inputs, proceedings of the Association for Computational Linguistics (ACL)","volume":"2024","author":"Jin","year":"2024","journal-title":"arXiv"},{"key":"ref15","doi-asserted-by":"publisher","first-page":"1472","DOI":"10.15394\/jdfsl.2017.1472","article-title":"File type identification using computational intelligence for digital forensics","volume":"1","author":"Karampidis","year":"2017","journal-title":"IEEE Trans. Inf. Forensics Secur."},{"key":"ref16","doi-asserted-by":"publisher","first-page":"6769","DOI":"10.18653\/v1\/2020.emnlp-main.550","article-title":"Dense passage retrieval for open-domain question answering proceedings of EMNLP","volume":"2020","author":"Karpukhin","year":"2020","journal-title":"arXiv"},{"key":"ref17","doi-asserted-by":"publisher","first-page":"11401","DOI":"10.48550\/arXiv.2005.11401","article-title":"Retrieval-augmented generation for knowledge-intensive NLP tasks, advances in neural information processing systems (NeurIPS)","volume":"2020","author":"Lewis","year":"2020","journal-title":"arXiv"},{"key":"ref18","author":"Moskewicz","year":"2001"},{"key":"ref19","doi-asserted-by":"publisher","first-page":"21654","DOI":"10.48550\/arXiv.2603.21654","article-title":"Towards secure retrieval-augmented generation: a comprehensive review of threats, defenses and benchmarks","volume":"2026","author":"Mu","year":"2026","journal-title":"arXiv"},{"key":"ref20","doi-asserted-by":"publisher","first-page":"585","DOI":"10.18653\/v1\/2024.acl-long.585","article-title":"RAGTruth: a hallucination Corpus for developing trustworthy retrieval-augmented language models, proceedings of ACL","volume":"2024","author":"Niu","year":"2024","journal-title":"arXiv"},{"key":"ref21","doi-asserted-by":"publisher","first-page":"8302","DOI":"10.48550\/arXiv.2306.08302","article-title":"Unifying large language models and knowledge graphs: a roadmap","volume":"2024","author":"Pan","year":"2024","journal-title":"IEEE Trans. Knowl. Data Eng."},{"key":"ref22","doi-asserted-by":"publisher","first-page":"6213","DOI":"10.47611\/jsrhs.v12i4.6213","article-title":"A retrieval-augmented generation based large language model benchmarked on a novel dataset","volume":"12","author":"Pichai","year":"2023","journal-title":"J. Stud. Res."},{"key":"ref23","doi-asserted-by":"publisher","first-page":"37","DOI":"10.5121\/ijnlc.2024.13103","article-title":"RAG-FUSION: a new take on retrieval-augmented generation","volume":"13","author":"Rackauckas","year":"2024","journal-title":"Int. J. Nat. Lang. Comput."},{"key":"ref24","doi-asserted-by":"publisher","first-page":"04568","DOI":"10.48550\/arXiv.2601.04568","article-title":"Neurosymbolic retrievers for retrieval-augmented generation","volume":"2026","author":"Saxena","year":"2026","journal-title":"arXiv"},{"key":"ref25","doi-asserted-by":"publisher","first-page":"16035","DOI":"10.48550\/arXiv.2506.16035","article-title":"Vision-guided chunking is all you need: enhancing retrieval-augmented generation with multimodal document understanding","volume":"2025","author":"Tripathi","year":"2025","journal-title":"arXiv"},{"key":"ref26","doi-asserted-by":"publisher","first-page":"10419","DOI":"10.48550\/arXiv.2508.10419","article-title":"Comorag: a cognitive-inspired memory-organized retrieval-augmented generation framework for stateful long-narrative reasoning","volume":"2025","author":"Wang","year":"2025","journal-title":"arXiv"},{"key":"ref27","doi-asserted-by":"publisher","first-page":"01959","DOI":"10.48550\/arXiv.2508.01959","article-title":"Sitemb-v1.5: improved context-aware dense retrieval for semantic association and long-document comprehension","volume":"2025","author":"Wu","year":"2025","journal-title":"arXiv"},{"key":"ref28","doi-asserted-by":"publisher","first-page":"03629","DOI":"10.48550\/arXiv.2210.03629","article-title":"ReAct: synergizing reasoning and acting in language models proceedings of ICLR","volume":"2023","author":"Yao","year":"2023","journal-title":"arXiv"},{"key":"ref29","doi-asserted-by":"publisher","first-page":"338","DOI":"10.1016\/S0019-9958(65)90241-0","article-title":"Fuzzy sets","volume":"8","author":"Zadeh","year":"1965","journal-title":"Inf. Control."},{"key":"ref30","author":"Zheng","year":"2026"}],"container-title":["Frontiers in Artificial Intelligence"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2026.1922508\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,17]],"date-time":"2026-09-17T05:33:58Z","timestamp":1789623238000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2026.1922508\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,17]]},"references-count":30,"alternative-id":["10.3389\/frai.2026.1922508"],"URL":"https:\/\/doi.org\/10.3389\/frai.2026.1922508","relation":{},"ISSN":["2624-8212"],"issn-type":[{"value":"2624-8212","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,17]]},"article-number":"1922508"}}