路径张量分解的知识图谱推理算法

发布时间：2018-01-26 02:24

本文关键词： 张量分解路径推理知识图谱路径排列算法(PRA)　出处：《模式识别与人工智能》2017年05期 　论文类型：期刊论文

【摘要】：现有张量分解技术在用于知识图谱学习和推理过程中时,只考虑知识图谱中实体与实体间的直接关系,忽略知识图谱图形结构的特点.因此,文中提出基于路径张量分解的知识图谱推理算法(PRESCAL),利用路径排列算法(PRA)获得知识图谱中各实体对间的关系路径.然后对实体对间的关系路径进行张量分解,并在优化更新过程中采用交替最小二乘法.实验表明,在路径问题回答任务和实体链接预测任务中,PRESCAL可以取得较好的预测准确率.
[Abstract]:In the process of knowledge map learning and reasoning, the existing Zhang Liang decomposition technology only considers the direct relationship between entity and entity in knowledge map, and ignores the characteristics of graphic structure of knowledge map. A knowledge map reasoning algorithm based on path Zhang Liang decomposition is proposed in this paper. The path alignment algorithm is used to obtain the relationship path between the entities in the knowledge map. Then Zhang Liang decomposes the relationship path between the entity pairs. In the process of optimization and updating, the alternating least square method is used. The experimental results show that PRESCAL can obtain better prediction accuracy in path question answer task and entity link prediction task.
【作者单位】：福州大学数学与计算机科学学院;福州大学福建省网络计算与智能信息处理重点实验室;厦门理工学院计算机与信息工程学院;
【基金】：国家自然科学基金项目青年基金项目(No.61300105) 福建省自然科学基金项目(No.2017J01755) 福建省引导性项目(No.2016Y0060,2014Y0005) 福建省教育厅科技项目(No.JA15082,JA14243);福建省教育厅中青年教师教育科研项目(No.JAT160077) 厦门留学人员科研项目(No.XRS201631401)资助~~
【分类号】：TP18
【正文快照】： Supported by Young Scientists Fund of National Natural Science Foundation of China(No.61300105),Natural Science Foundation ofFujian Province(No.2017J01755),Science Foundation of Fujian Province(No.2016Y0060,2014Y0005),Science and TechnologyProject of the

【参考文献】