海棠木瓜腊八粥教案第45卷第2期2020年4月 广西大学学报(自然科学版)Journal of Guangxi University(Nat Sci Ed)Vol.45No.2Apr.2020 收稿日期:2019⁃11⁃25；修订日期:2019⁃12⁃25 基金项目：天津市科技计划项目(18ZXZNGX00370)；国家自然科学基金资助项目(61572350) 通讯作者：曹扬(1981 )，男，重庆巫山人，中电科大数据研究院高级工程师;E⁃mail:㊂
引文格式：昌攀，曹扬.改进的TransH 模型在知识表示与推理领域的研究[J].广西大学学报(自然科学版),2020,45(2):321⁃327.doi:10.13624/jki.issn.1001⁃7445.2020.0321改进的TransH 模型在知识表示与推理领域的研究
昌攀，曹扬*
(中电科大数据研究院国家工程实验室，贵州贵阳550081)
摘要：针对Trans 系列的知识图谱表示与推理模型在训练的过程中，随机构造正负例三元组样本进行训练，没有考虑替换的实体与原实体之间存在的相似度差异度关系，导致模型无法识别实体之间的相似度，效果低下㊂在TransH 模型的构建基础上，采用单层神经网络的非线性操作来精确刻画实体和关系之间的语义信息，同时创新性地加入了正㊁负三元组之间的头/尾实体之间的差异度信息，用于校正正㊁负三元组样本之间的联系，使模型能够辨别替换的实体与原实体间的相似度，进而提出了m TransH 模型㊂实验证明:m TransH 模型在知识图谱的链接预测任务中，提高了模型对正例样本的辨识度，从而提高知怪味花生的做法
识推理的链接预测准确率㊂青香蕉苹果
关键词：知识图谱；表示；推理；相似度;m TransH 模型；链接预测
幼儿园入园
中图分类号:TP391 文献标识码:A 文章编号:1001⁃7445(2020)02⁃0321⁃07用晴组词
美丽的遇见Study on a modified TransH⁃bad model in knowledge reprentation and reasoning field
CHANG Pan,CAO Yang *
(National Engineering Laboratory for CETC Rearch Institute of Big Data,Guiyang 550081,China)
抓耳挠腮Abstract :During the training of knowledge graph reprentation and reasoning model for Trans ries,the relationship of similarity and difference between the replaced head or tail entity and the original entity are not considered when positive and negative sample triples are constructed for training,which leads to low efficiency and the model’s inability to identify the similarity between entities.In this study,bad on the construction of the TransH model,the nonlinear operation of single layer neural network is ud to describe the mantic information between entity and relation.As an innovation,the difference between the head and tail entities of the positive and negative triples is introduced,so as to correct the relationship between the positive and negative triplet samples and to
enable the m TransH model to identify the similarity between the replaced entity and the original entity,and then the m TransH model is propod.Experimental results show that the m TransH model improves the recognition of positive samples in the link prediction task of knowledge graph,thus improving the link prediction accuracy of knowledge reasoning.
