Reproduce “LLM-Guided Ontology-Driven Knowledge Graph Construction from Unstructured Text” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Abdelhadi Belfadel, Maxence Gagnant, Joseph Kattan, Sana Tmar
OntoConnectLM uses locally deployable open language models to extract typed entities and relationship triples from incident reports, then feed them into ontology construction. On 80 manually annotated French power-grid reports, schema-rich prompting raised Gemma3 27B strict NER F1 from 0.42 to 0.57 and exact-match relation F1 from 0.51 to 0.70. These scores measure extraction, not completed graph quality or generalization.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.