现代信息科技2025，Vol.9Issue(16)：44-49,56,7.DOI:10.19850/j.cnki.2096-4706.2025.16.009
基于机器阅读理解的水利工程巡检知识抽取
Knowledge Extraction for Water Conservancy Engineering Inspection Based on Machine Reading Comprehension
翟向超 1张健豪 1韩文豪1
作者信息
- 1. 华北水利水电大学 信息工程学院,河南 郑州 450046
- 折叠
摘要
Abstract
The inspection data of water conservancy projects contains rich risk information.Aiming at the problems of complex long entities and nested entities in the data,a hierarchical ontology model and a multi-task knowledge extraction framework are constructed to solve the problem of entity and relationship extraction.Firstly,aiming at the complexity of domain data,the hierarchical strategy is used to effectively solve the modeling problem of complex inspection data.Secondly,a knowledge extraction model combining machine reading comprehension and multi-task learning is proposed innovatively.The model includes entity extraction task based on question answering,entity reclassification task based on description discrimination and relationship extraction task.Each task realizes collaborative optimization through shared parameters and joint training.Finally,experimental verification shows that the entity and relationship extraction effect of the proposed method is significantly better than other baseline models,and it can meet the actual needs of knowledge extraction of engineering inspection data.关键词
水利工程巡检数据/知识抽取/机器阅读理解/多任务学习Key words
water conservancy project inspection data/knowledge extraction/machine reading comprehension/multi-task learning分类
信息技术与安全科学引用本文复制引用
翟向超,张健豪,韩文豪..基于机器阅读理解的水利工程巡检知识抽取[J].现代信息科技,2025,9(16):44-49,56,7.
