New LLM pipeline automates ontology generation from databases

Researchers have developed RIGOR, a new LLM-driven pipeline designed to automatically generate OWL ontologies from relational database schemas. This system aims to improve semantic interoperability and facilitate downstream tasks like knowledge graph population and automated reasoning. RIGOR generates direct mappings for schema coverage and then enriches them by retrieving information from the schema context, external ontologies, and an incrementally growing ontology. A Gen-LLM creates ontology fragments, which are validated by a separate Judge-LLM before integration, with the process iterating until the entire schema is covered. Experiments show RIGOR outperforms existing methods without human intervention. AI

IMPACT Automates ontology creation, potentially accelerating knowledge graph population and semantic interoperability in data management.

RANK_REASON The cluster describes a new research paper detailing a novel method for generating ontologies from relational databases using LLMs. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• alphaXiv
• arXiv
• Gen-LLM
• Hugging Face
• Judge-LLM
• Mojtaba Nayyeri
• OWL2DL
• RIGOR

• paper
• product

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