火力与指挥控制2025，Vol.50Issue(7)：18-26,9.DOI:10.3969/j.issn.1002-0640.2025.07.003
基于知识图谱增强的文本融合问答推理方法
Text Fusion Question-answering Inference Method Based on Knowledge Graph Enhancement
摘要
Abstract
A text fusion question-answering inference model GAGN based on knowledge graph en-hancement is proposed.By combining knowledge graph and text data,the missing knowledge graph data is supplemented,and the information of knowledge graph is updated by graph attention neural network,which constructs heterogeneous sub-graphs with text and performs question-answering inference.GAGN model can effectively solve the problem of inaccurate question-answering inference caused by poor data integrity in the field of military question-answering,and can promote the construction process of artificial intelligence military staff.The GAGN model is tested on two general question-answering datasets and one military question-answering dataset,and the superiority of GAGN model is verified.关键词
问答系统/知识图谱/图注意力神经网络/异构子图/问答推理Key words
question-answering system/knowledge graph/graph attention neural network/heteroge-neous sub-graphs/question-answering inference分类
信息技术与安全科学引用本文复制引用
姚奕,陈朝阳,尹瑞江,张帆,霍炎..基于知识图谱增强的文本融合问答推理方法[J].火力与指挥控制,2025,50(7):18-26,9.基金项目
国家自然科学基金(62273356,61806221) （62273356,61806221）
高层次科技创新人才自主科研基金资助项目(KYZYJKKC0024001) （KYZYJKKC0024001）
