Retrieval-Augmented Generation of Ontologies from Relational Databases

2025-06-02

AI Summary

To address the challenges of ontology scarcity, high manual effort, and poor interoperability in transforming relational databases into high-semantic-quality knowledge graphs, this paper proposes RIGOR: a retrieval-augmented, iterative ontology generation framework. RIGOR integrates database schemas, domain ontology repositories, and a dynamically evolving core ontology, leveraging foreign-key-driven iterations and a dual-LLM coordination mechanism—where a generative LLM constructs OWL ontologies and a discriminative LLM ensures semantic consistency and formal verifiability. This enables traceable, incremental, and fully automated ontology engineering. Experiments on real-world databases demonstrate that RIGOR-generated ontologies significantly outperform baseline approaches in accuracy, completeness, and logical consistency, while reducing human intervention by over 70%. The resulting ontologies effectively support semantic interoperability and graph neural network–based reasoning.

Technology Category

• Knowledge Representation and Reasoning: Ontologies
• Data Mining & Knowledge Management: Linked Open Data, Knowledge Graphs & KB Completion
• Machine Learning: Statistical Relational/Logic Learning

Application Category

• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

Transforming relational databases into knowledge graphs with enriched ontologies enhances semantic interoperability and unlocks advanced graph-based learning and reasoning over data. However, previous approaches either demand significant manual effort to derive an ontology from a database schema or produce only a basic ontology. We present RIGOR, Retrieval-augmented Iterative Generation of RDB Ontologies, an LLM-driven approach that turns relational schemas into rich OWL ontologies with minimal human effort. RIGOR combines three sources via RAG, the database schema and its documentation, a repository of domain ontologies, and a growing core ontology, to prompt a generative LLM for producing successive, provenance-tagged delta ontology fragments. Each fragment is refined by a judge-LLM before being merged into the core ontology, and the process iterates table-by-table following foreign key constraints until coverage is complete. Applied to real-world databases, our approach outputs ontologies that score highly on standard quality dimensions such as accuracy, completeness, conciseness, adaptability, clarity, and consistency, while substantially reducing manual effort.

Problem

• Automating ontology generation from relational databases
• Enhancing semantic interoperability with enriched ontologies
• Reducing manual effort in creating comprehensive OWL ontologies

Innovation

• LLM-driven RDB to OWL ontology conversion
• RAG combines schema, docs, domain ontologies
• Iterative judge-LLM refinement of delta fragments