Reproduce “Using Large Language Models for OntoClean-based Ontology Refinement” 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
Yihang Zhao, Neil Vetter, Kaveh Aryan
This research explores the application of Large Language Models (LLMs) to automate the rigorous OntoClean methodology for ontology refinement. The study demonstrates that advanced LLMs, particularly GPT-4, can accurately assign OntoClean meta-properties to ontology classes, achieving nearly perfect performance when provided with hierarchical representations of the ontology.
Key insights
- GPT-4 exhibits near-perfect accuracy in assigning OntoClean meta-properties, outperforming GPT-3.5 significantly.
- Hierarchical ontology representations are crucial for LLMs to accurately assign OntoClean meta-properties; flat representations lead to poor performance.
- For advanced LLMs like GPT-4, zero-shot prompting is sufficient, indicating they possess inherent knowledge of OntoClean principles without needing extensive in-context definitions.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.