Integrating Ontology and LLMs for Diagnosis and Repair of Concrete Surface Defects
Open Access
- 27 July 2025
- conference paper
- Published by International Association for Automation and Robotics in Construction (IAARC) in Proceedings of the 37th International Symposium on Automation and Robotics in Construction (isarc)
- Vol. 2025 Proce, 1292-1299
Abstract
This paper explores the integration of a novel Ontology for Concrete Surface Defects (OCSD) with a Large Language Model (LLM), specifically GPT-4o (omni), to enhance defect diagnosis and repair strategies in concrete structures. While LLMs independently offer significant reasoning and natural language capabilities, this study demonstrates the value of combining their interpretative power with the structured knowledge representation provided by OCSD. By enabling adaptive reasoning, where the LLM relies on the ontology's domain-specific relationships and thresholds, and updates its conclusions for specific defect types, the system flexibly adjusts its decision-making based on context. This integration improves diagnosis accuracy, reasoning transparency, and decision-making efficiency. The proposed method is validated through a case study, highlighting the synergy of OCSD and GPT-4o in addressing challenges in defects diagnosis and repair.Keywords
This publication has 8 references indexed in Scilit:
- OntoChat: A Framework for Conversational Ontology Engineering Using Language Models2025
- Developing an Ontology for Concrete Surface Defects to Enhance Inspection, Diagnosis and Repair Information ModelingInfrastructures, 2024
- Ontology-assisted GPT-based building performance simulation and assessment: Implementation of multizone airflow simulationEnergy and Buildings, 2024
- Combining Ontological Knowledge and Large Language Model for User-Friendly Service Robots2024
- Towards Next-Generation Urban Decision Support Systems through AI-Powered Construction of Scientific Ontology Using Large Language Models—A Case in Optimizing Intermodal Freight TransportationSmart Cities, 2024
- Ontology-integrated tuning of large language model for intelligent maintenanceCIRP Annals, 2024
- Dynamic graph CNN based semantic segmentation of concrete defects and as-inspected modelingAutomation in Construction, 2024
- OntoChatGPT Information System: Ontology-Driven Structured Prompts for ChatGPT Meta-LearningInternational Journal of Computing, 2023