Ontology-integrated tuning of large language model for intelligent maintenance

IF 3.2 3区工程技术 Q2 ENGINEERING, INDUSTRIAL

Cirp Annals-Manufacturing Technology Pub Date : 2024-01-01 DOI:10.1016/j.cirp.2024.04.012

引用次数: 0

Abstract

As new AI technologies such as Large Language Models (LLM) quickly evolve, the need for enhancing general-purpose LLMs with physical knowledge to better serve the manufacturing community has been increasingly recognized. This paper presents a method that tailors GPT-3.5 with domain-specific knowledge for intelligent aircraft maintenance. Specifically, aircraft ontology is investigated to curate maintenance logs with encoded component hierarchical structure to fine-tune GPT-3.5. Experimental results demonstrate the effectiveness of the developed method in accurately identifying defective components and providing consistent maintenance action recommendations, outperforming general-purpose GPT-3.5 and GPT-4.0. The method can be adapted to other domains in manufacturing and beyond.

查看原文本刊更多论文

本体集成调整大型语言模型，实现智能维护

随着大型语言模型（LLM）等新型人工智能技术的快速发展，人们越来越认识到需要利用物理知识增强通用 LLM，以便更好地服务于制造业。本文介绍了一种利用特定领域知识为飞机智能维护量身定制 GPT-3.5 的方法。具体来说，本文研究了飞机本体，通过编码组件分层结构来整理维护日志，从而对 GPT-3.5 进行微调。实验结果表明，所开发的方法在准确识别缺陷部件和提供一致的维护行动建议方面非常有效，优于通用的 GPT-3.5 和 GPT-4.0。该方法可适用于制造业及其他领域。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Cirp Annals-Manufacturing Technology 工程技术-工程：工业

CiteScore

7.50

自引率

9.80%

发文量

137

审稿时长

13.5 months

期刊介绍： CIRP, The International Academy for Production Engineering, was founded in 1951 to promote, by scientific research, the development of all aspects of manufacturing technology covering the optimization, control and management of processes, machines and systems. This biannual ISI cited journal contains approximately 140 refereed technical and keynote papers. Subject areas covered include: Assembly, Cutting, Design, Electro-Physical and Chemical Processes, Forming, Abrasive processes, Surfaces, Machines, Production Systems and Organizations, Precision Engineering and Metrology, Life-Cycle Engineering, Microsystems Technology (MST), Nanotechnology.