LLM pipeline translates institutional policies into machine-readable knowledge graphs

Researchers have developed PolicyKG, an agentic LLM pipeline designed to convert institutional policies from natural language into machine-readable SHACL knowledge graphs. This system classifies policy sentences into obligations, permissions, or prohibitions, translates them into first-order deontic logic, and generates SHACL constraints. The pipeline utilizes a LangGraph state machine and a Corpus Adapter for domain retargeting, demonstrating high accuracy on institutional policy corpora and showing promise for compliance checking. AI

IMPACT Enables automated compliance checking and policy management by converting unstructured text into structured, machine-readable formats.

RANK_REASON The item is an academic paper detailing a novel LLM pipeline for a specific research task. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• Asian Institute of Technology
• Corpus Adapter
• General Data Protection Regulation
• LangGraph
• LexDeMod
• PolicyKG
• SHACL

• paper
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →