Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
The paper introduces De Jure, an automated pipeline for extracting structured regulatory rules from raw documents using LLMs, eliminating the need for human annotation or domain-specific prompting. De Jure normalizes documents, decomposes them into rule units, evaluates extractions across 19 dimensions using an LLM-as-a-judge, and iteratively repairs low-scoring extractions. Evaluated on finance, healthcare, and AI governance corpora, De Jure achieves high extraction quality and significantly improves downstream compliance question-answering performance compared to prior work.
Forget hand-annotated data: De Jure shows LLMs can iteratively self-refine to extract structured regulatory rules from raw documents, guided by explicit, interpretable evaluation criteria.
Regulatory documents encode legally binding obligations that LLM-based systems must respect. Yet converting dense, hierarchically structured legal text into machine-readable rules remains a costly, expert-intensive process. We present De Jure, a fully automated, domain-agnostic pipeline for extracting structured regulatory rules from raw documents, requiring no human annotation, domain-specific prompting, or annotated gold data. De Jure operates through four sequential stages: normalization of source documents into structured Markdown; LLM-driven semantic decomposition into structured rule units; multi-criteria LLM-as-a-judge evaluation across 19 dimensions spanning metadata, definitions, and rule semantics; and iterative repair of low-scoring extractions within a bounded regeneration budget, where upstream components are repaired before rule units are evaluated. We evaluate De Jure across four models on three regulatory corpora spanning finance, healthcare, and AI governance. On the finance domain, De Jure yields consistent and monotonic improvement in extraction quality, reaching peak performance within three judge-guided iterations. De Jure generalizes effectively to healthcare and AI governance, maintaining high performance across both open- and closed-source models. In a downstream compliance question-answering evaluation via RAG, responses grounded in De Jure extracted rules are preferred over prior work in 73.8% of cases at single-rule retrieval depth, rising to 84.0% under broader retrieval, confirming that extraction fidelity translates directly into downstream utility. These results demonstrate that explicit, interpretable evaluation criteria can substitute for human annotation in complex regulatory domains, offering a scalable and auditable path toward regulation-grounded LLM alignment.