Exploring Large Language Models for Ontology Alignment

TL;DR

This paper explores the use of large generative language models like GPT-3.5 and Flan-T5 for ontology alignment, demonstrating their potential to outperform existing systems in identifying concept equivalences across ontologies.

Contribution

It is the first to evaluate zero-shot performance of LLMs on ontology alignment tasks, highlighting their potential advantages over traditional methods.

Findings

LLMs can outperform BERTMap with proper prompt design

Zero-shot performance of LLMs is promising for ontology matching

Preliminary results show potential for LLMs in ontology alignment

Abstract

This work investigates the applicability of recent generative Large Language Models (LLMs), such as the GPT series and Flan-T5, to ontology alignment for identifying concept equivalence mappings across ontologies. To test the zero-shot performance of Flan-T5-XXL and GPT-3.5-turbo, we leverage challenging subsets from two equivalence matching datasets of the OAEI Bio-ML track, taking into account concept labels and structural contexts. Preliminary findings suggest that LLMs have the potential to outperform existing ontology alignment systems like BERTMap, given careful framework and prompt design.