Ontologies are useful structures for organizing and maintaining information that can be understood by both humans and systems. However, the manual crafting of ontologies is labor-intensive, leading to a lack of reference ontologies in many specific domains. The exceptional natural language understanding capabilities of Large Language Models (LLMs) have prompted their application in various fields, including ontology development. This work presents an experimental technique using LLMs as domain experts to construct conceptual hierarchies for a given initial concept. Twenty ontologies were automatically constructed for the Brazilian maritime territory (a.k.a. the Blue Amazon) using GPT-3.5 and GPT-4, and evaluated by human experts. The models were able to create overall coherent conceptualizations of the domain, but none of the outputs were completely satisfactory as representations of the context without further refinement.
Blogger's Review: This paper highlights the potential of LLMs in constructing domain ontologies. Although the results require improvement, the value of LLMs as preliminary tools is undeniable. Future research could explore human-machine collaboration to further enhance the quality and accuracy of ontologies.