POLICYIQ: AN AI-POWERED POLICY CONFLICT VERIFICATION SYSTEM | IJCSE Volume 10 β Issue 5 | IJCSE-V10I5P24
Table of Contents
ToggleInternational Journal of Computer Science Engineering Techniques
ISSN: 2455-135X
Volume 10, Issue 5
|
Published:
Author
KOPPOLU SAHITH, VADDE VENKI, VIPUL PONUGOTI, KORADA VENKATA KARTHIK, KORADA RAMESH
Abstract
Organisations, governments and regulated industries operate under large, continuously revised bodies of policy. As these documents evolve across versions, editions and jurisdictions, overlapping, contradictory or silently modified clauses accumulate and create compliance, governance and legal risk that manual review cannot scale to detect. To resolve these systemic bottlenecks, this paper presents PolicyIQ, an enterprise-grade artificial intelligence platform for automated policy conflict verification and version-aware change intelligence. PolicyIQ ingests heterogeneous policy documents through a robust extraction pipeline with optical character recognition (OCR) fallback, decomposes them into atomic clauses using a hybrid rule-based and semantic segmentation engine, and encodes each clause into a 384-dimensional dense vector using a Sentence-BERT transformer (MiniLM-L6). A cross-version cosine-similarity matrix drives a multi-stage classifier that labels every clause transition as duplicate, modified, new, removed or direct-conflict, augmented by deontic, numeric and negation-scope heuristics for typed contradiction detection. A distribution-aware adaptive thresholding scheme replaces brittle fixed cut-offs by deriving decision boundaries from the per-corpus similarity distribution. Built on a decoupled four-tier microservices architecture utilising a FastAPI security gateway, asynchronous Celery workers, PostgreSQL relational storage and a Qdrant/FAISS vector store, PolicyIQ delivers explainable dual-audience summaries while maintaining data privacy through JWT authentication and role-based access control. On a controlled synthetic policy corpus the platform attains target performance of 91.6% precision for adaptive conflict classification and 0.944 segmentation F1, with sub-second clause-level inference, demonstrating a reproducible, provenance-aware pathway from raw policy text to actionable conflict intelligence.
Keywords
PolicyIQ, Artificial Intelligence, Policy Conflict Verification, Semantic Textual Similarity, Sentence-BERT, Clause Segmentation, Adaptive Thresholding, Contradiction Detection, Vector Search, Microservices Architecture, Legal NLP, Version Diffing, Explainable Summarisation.Conclusion
This paper presented PolicyIQ, a version-aware, multi-stage artificial intelligence platform for automated policy conflict verification and change intelligence. By unifying robust text extraction, hybrid clause segmentation, dense Sentence-BERT embeddings, cross-version typed conflict classification, distribution-aware adaptive thresholding and dual-audience summarisation within a hardened four-tier microservices architecture, PolicyIQ demonstrates a reproducible pathway from raw policy text to explainable, actionable conflict intelligence. The design directly targets the gaps left by lexical diffing and version-blind similarity tools, replacing opaque scalar scores with typed, provenance-preserving verdicts secured behind an authenticated, role-based API. On a controlled synthetic benchmark the platform meets its design targets for segmentation and adaptive conflict classification while sustaining sub-second clause-level inference, establishing a solid foundation for rigorous, human-annotated evaluation.
References
[1] T. Mikolov, K. Chen, G. Corrado, and J. Dean, βEfficient estimation of word representations in vector space,β in Proc. ICLR Workshop, 2013.
[2] A. Vaswani et al., βAttention is all you need,β in Proc. NeurIPS, Long Beach, CA, USA, 2017, pp. 5998β6008.
[3] J. Devlin, M.-W. Chang, K. Lee, and K. Toutanova, βBERT: Pre-training of deep bidirectional transformers for language understanding,β in Proc. NAACL-HLT, 2019, pp. 4171β4186.
[4] N. Reimers and I. Gurevych, βSentence-BERT: Sentence embeddings using Siamese BERT-networks,β in Proc. EMNLP-IJCNLP, 2019, pp. 3982β3992.
[5] S. R. Bowman, G. Angeli, C. Potts, and C. D. Manning, βA large annotated corpus for learning natural language inference,β in Proc. EMNLP, 2015, pp. 632β642.
[6] Y. Koreeda and C. D. Manning, βContractNLI: A dataset for document-level natural language inference for contracts,β in Findings of EMNLP, 2021, pp. 1907β1919.
[7] LegalWiz, βA multi-agent generation framework for contradiction detection in legal documents,β arXiv preprint arXiv:2510.03418, 2025.
[8] βNLP-based regulatory compliance: Using GPT-4.0 to decode regulatory documents,β arXiv preprint arXiv:2412.20602, 2024.
[9] M. Lippi et al., βDeep learning for conflicting statements detection in text,β PeerJ Preprints, vol. 6, e26589v1, 2018.
[10] S. Gupta et al., βIdentifying contradictions in the legal proceedings using natural language models,β SN Computer Science, vol. 3, no. 3, 2022.
[11] A. Mourad and H. Jebbaoui, βConflict detection and resolution of XACML policies,β in Proc. IEEE Symp. Computers and Communications, 2014.
[12] M. A. Rahman et al., βClustering-based approach for anomaly detection in XACML policies,β in Proc. SECRYPT, 2017, pp. 548β553.
[13] H. Shah and P. Kumar, βAutomated interpretation of regulations using NLP for compliance-centric analysis of legal texts,β ResearchGate preprint, 2024.
[14] R. Mohan et al., βDetecting compliance of privacy policies with data protection laws,β arXiv preprint arXiv:2102.12362, 2021.
[15] D. Silva et al., βAnalysing similarities between legal court documents using NLP approaches based on transformers,β PLoS ONE / PMC, 2024.
[16] I. Chalkidis et al., βLEGAL-BERT: The muppets straight out of law school,β in Findings of EMNLP, 2020, pp. 2898β2904.
[17] J. Johnson, M. Douze, and H. JΓ©gou, βBillion-scale similarity search with GPUs,β IEEE Trans. Big Data, vol. 7, no. 3, pp. 535β547, 2021.
[18] Yu. A. Malkov and D. A. Yashunin, βEfficient and robust approximate nearest neighbor search using hierarchical navigable small world graphs,β IEEE Trans. Pattern Anal. Mach. Intell., vol. 42, no. 4, pp. 824β836, 2020.
[19] R. Mihalcea and P. Tarau, βTextRank: Bringing order into texts,β in Proc. EMNLP, 2004, pp. 404β411.
[20] R. S. Sandhu, E. J. Coyne, H. L. Feinstein, and C. E. Youman, βRole-based access control models,β IEEE Computer, vol. 29, no. 2, pp. 38β47, 1996.
[21] C. D. Manning, P. Raghavan, and H. SchΓΌtze, Introduction to Information Retrieval. Cambridge, U.K.: Cambridge Univ. Press, 2008.
[22] R. Smith, βAn overview of the Tesseract OCR engine,β in Proc. ICDAR, 2007, pp. 629β633.
[23] Y. Liu et al., βRoBERTa: A robustly optimized BERT pretraining approach,β arXiv preprint arXiv:1907.11692, 2019.
[24] W. Wang et al., βMiniLM: Deep self-attention distillation for task-agnostic compression of pre-trained transformers,β in Proc. NeurIPS, 2020.
[25] S. RamΓrez, βFastAPI: Modern, fast web framework for building APIs with Python,β software documentation, 2023.
[26] Qdrant Team, βQdrant: Vector similarity search engine,β software documentation, 2024.
