Researchers have developed a two-stage workflow to assist in the engineering and construction of a French legal knowledge graph, specifically focusing on maintenance regulations. This process involves open extraction of entities and triples, normalization of labels using embedding-based fusion, and induction of candidate object properties. Subsequent stages utilize the generated ontology for closed extraction and RDF graph construction across the full corpus. Experiments using GPT-4.1 and mistral-large-2512 demonstrated effective structured outputs and class alignment, significantly reducing duplicated entities and predicates. AI
IMPACT This research demonstrates a novel application of LLMs for complex legal text analysis, potentially streamlining the creation of structured legal knowledge bases.
RANK_REASON The cluster contains an academic paper detailing a novel methodology for ontology engineering and knowledge graph construction using LLMs. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
