Leveraging Large Language Models for Ontology Requirements Engineering

EasyChair Preprint 15963

Abstract

Ontologies are essential for structuring domain knowledge, enabling shared understanding to address the challenges of exponential web data growth. Ontology Engineering (OE) has evolved into a collaborative, community-driven practice, with Ontology Requirements Engineering (ORE) providing a systematic framework for capturing, documenting, and validating requirements to support ontology development, evaluation, and maintenance. However, ORE still relies on manual techniques such as brainstorming, interviews, and spreadsheets, making the process resource-intensive. Recent advances in Large Language Models (LLMs) present new opportunities to support ORE tasks. Existing studies highlight their potential in ontology user story generation, as well as competency questions (CQs) generation and retrofitting. However, LLM-based ORE frameworks are still in their early stages and lack structured guidance across the full ORE workflow. Therefore, this research aims to bridge the gap by investigating how ORE tasks can be potentially supported by LLMs and developing the conversational agent OntoChat to integrate LLMs for assisting users in these tasks. In this paper, we present preliminary findings on how LLMs can potentially support ORE based on the first year of this research.

Keyphrases: Competency Questions, LLMs, Ontology Engineering, Requirements Engineering, User Stories, large language models

@booklet{EasyChair:15963,
  author    = {Yihang Zhao},
  title     = {Leveraging Large Language Models for Ontology Requirements Engineering},
  howpublished = {EasyChair Preprint 15963},
  year      = {EasyChair, 2025}}