摘要
随着人工智能发展,大模型作为新质生产力重塑特种设备检验行业。本文针对特种设备检验业务复杂性,首先概述了大模型在特种设备检验机构中的应用现状,然后设计了包含数据采集层、数据预处理层、模型层、应用层的框架,集成多模态数据融合、知识图谱构建、小样本学习、人机协同优化等关键技术,接着给出了智能问答、检验方案协助制定、辅助诊断推理、风险监测与预警等典型应用场景,最后给出了大模型在提升检验机构智能化水平的同时需解决数据安全、算法可解释性等问题的建议,为特种设备检验机构智能化转型提供技术支撑。
With the development of artificial intelligence,large model,as a new form of productive force,is reshaping the special equipment inspection industry.Aiming at the complexity of special equipment inspection,this paper first overviews the application status of large model in special equipment inspection institutions,then designs a framework including data layer,preprocessing layer,model layer and application layer,integrates key technologies such as multimodal data fusion,knowledge graph,few-shot learning and human-machine collaborative optimization,then gives typical application scenarios such as intelligent question answering,inspection plan formulation,auxiliary diagnosis and risk monitoring.Finally it puts forward suggestions on how large model can improve the intelligence level of inspection institutions while solving problems such as data security and model interpretability,so as to provide technical support for the intelligent transformation of special equipment inspection institutions.
作者
胡静波
朱博文
王会方
HU Jingbo;ZHU Bowen;WANG Huifang(Nanjing Special Equipment Safety Supervision Inspection and Research Institute,Nanjing,210017)
出处
《中国特种设备安全》
2025年第S1期116-121,共6页
China Special Equipment Safety
关键词
大模型
特种设备检验
多模态融合
知识图谱
Large model
Special equipment inspection
Multimodal fusion
Knowledge graph
