一种基于改进向量投影距离的知识图谱表示方法

doi:10.11896/jsjkx.190300024

计算机科学 ›› 2020, Vol. 47 ›› Issue (4): 189-193.doi: 10.11896/jsjkx.190300024

一种基于改进向量投影距离的知识图谱表示方法

李鑫超, 李培峰, 朱巧明

苏州大学计算机科学与技术学院江苏苏州215006江苏省计算机信息处理技术重点实验室江苏苏州215006

收稿日期:2019-03-08 出版日期:2020-04-15 发布日期:2020-04-15
通讯作者: 李培峰(pfli@suda.edu.cn)
基金资助:
国家自然科学基金项目(61836007,61772354,61773276)

Knowledge Graph Representation Based on Improved Vector Projection Distance

LI Xin-chao, LI Pei-feng, ZHU Qiao-ming

School of Computer Science and Technology,Soochow University,Suzhou,Jiangsu 215006,ChinaProvincial Key Laboratory for Computer Information Processing Technology,Suzhou,Jiangsu 215006,China

Received:2019-03-08 Online:2020-04-15 Published:2020-04-15
Contact: LI Pei-feng,born in 1971,Ph.D,professor,Ph.D supervisor,is a member of China Computer Federation.His main research interests include natural language processing and machine learning.
About author:LI Xin-chao,born in 1995,postgradua-te.His main research interests include natural language processing and representation learning.
Supported by:
This work was supported by the National Natural Science Foundation of China(61836007,61772354,61773276).

摘要/Abstract

摘要： 表示学习在知识图谱推理中有着重要的研究价值,将知识库中的实体和关系用连续低维向量进行表示,可实现知识的可计算。基于向量投影距离的知识表示学习模型在面对复杂关系时有较好的知识表达能力,但在处理一对一简单关系时容易受到无关信息的干扰,并且在一对多、多对一和多对多等复杂关系上存在性能提升空间。为此,文中提出了一个基于改进向量投影距离的知识表示学习模型SProjE,该模型引入自适应度量方法,降低了噪声信息的影响。在此基础上,通过进一步优化损失函数来提高复杂关系三元组的损失权重。该模型适用于大规模知识图谱的表示学习任务。最后,在标准知识图谱数据集WN18和FB15K上分析和验证了所提方法的有效性,基于链路预测任务的评测实验结果表明,相较于现有的模型和方法,SProjE在各项性能指标上均取得了明显的进步。

关键词: 知识图谱, 表示学习, 自适应度量, 链路预测

Abstract: Representation learning is of great value in knowledge graph reasoning,which realizes the computability of knowledge by embedding entities and relationships into a low-dimensional space.The representation learning model based on vector projection distance has better ability of knowledge representation on complex relationships.However,the model is easily susceptible to irrelevant information,especially when dealing with one-to-one relationships,and it still has space to improve performance in representing one-to-many,many-to-one and many-to-many relationships.In this paper,we proposed an improved representation learning model SProjE,which introduces an adaptive metric method to reduce the weight of noise information and optimizes the loss function to improve the loss weight of complex relation triples.The proposed model is suitable for large scale knowledge graph representation learning.At last,the experimental results on the WN18 and FB15k data sets show that SProjE achieves significant and consistent improvements compared with the existing models and methods.

Key words: Knowledge graph, Representation learning, Adaptive metric, Entity link prediction

中图分类号:

TP391.1

李鑫超, 李培峰, 朱巧明. 一种基于改进向量投影距离的知识图谱表示方法[J]. 计算机科学, 2020, 47(4): 189-193.

LI Xin-chao, LI Pei-feng, ZHU Qiao-ming. Knowledge Graph Representation Based on Improved Vector Projection Distance[J]. Computer Science, 2020, 47(4): 189-193.

参考文献

[1]BORDES A,WESTON J,USUNIER N.Open Question Answering with Weakly Supervised EmbeddingModels[J/OL].
[2014-04-16].https://arxiv.org/pdf/1404.4326.pdf.
[2]ZHENG Z,SI X,LI F,et al.Entity Disambiguation with Freebase[C]//2012 IEEE/WIC/ACM International Conferences on Web Intelligence and Intelligent Agent Technology.IEEE Computer Society,2012.
[3]DAIBER J,JAKOB M.Improving efficiency and accuracy inmultilingual entity extraction[C]//Proceedings of the 9th International Conference on Semantic Systems.ACM,2013:121-124.
[4]BERANT J,CHOU A,FROSTIG R,et al.Semantic parsing on freebase from question-answer pairs[C]//Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing.2013:1533-1544.
[5]WANG Q.Knowledge graph embedding:A survey of approaches and applications [J].IEEE Transactions on Knowledge and Data Engineering,2017,29(12):2724-2743.
[6]SHI B,WENINGER T.ProjE:Embedding Projection forKnowledge Graph Completion[C]// Proc of AAAI 2017.San Francisco:AAAI,2017:1236-1242.
[7]MIKOLOV T,SUTSKEVER I,CHEN,et al.Distributed representations of words and phrases and their compositionality[C]//Proc of NIPS 2013.Cambridge,MA:MIT Press,2013:3111-3119.
[8]BORDES A,USUNIER N,GARCIA-DURAN A,et al.Translating embeddings for modeling multi-relational data[C]//Proc of NIPS 2013.Cambridge,MA:MIT Press,2013:2787-2795.
[9]WANG Z,ZHANG J,FENG J,et al.Knowledge graph embedding by translating on hyperplanes[C]//Proc of AAAI 2014.Menlo Park,CA:AAAI,2014:1112-1119.
[10]LIN Y K,LIU Z Y,SUN M S,et al.LearingEntity and Relation Embeddings for Knowledge Graph Completion[C]//Proc of AAAI 2015.Menlo Park,CA:AAAI,2015:2181-2187.
[11]JI G,HE S,XU L,et al.Knowledge Graph Embedding viaDynamic Mapping Matrix[C]//Proc of ACL 2015.Beijing,China:ACL,2015:687-696.
[12]JI G,LIU K,HE S,et al.Knowledge Graph Completion withAdaptive Sparse Transfer Matrix[C]// Thirtieth Aaai Confe-rence on Artificial Intelligence.AAAI Press,2016.
[13]FAN M,QIANG Z,CHANG E,et al.Transition-based knowledge graph embedding with relational mapping properties[C]//Proc. of PACLIC.2014.
[14]XIAO H,HUANG M L,HAO Y,et al.TransA:An adaptiveapproach for knowledge graph embedding [J/OL].[2015-09-28].https://arxiv.org/pdf/1509.05490.pdf.
[15]NICKEL M,TRESP V,KRIEGEL H P.A Three-Way Modelfor Collective Learning on Multi-Relational Data[C]//Proc of ICML 2011.New York,USA:ACM,2011:809-816.
[16]BORDES A,GLOROT X,WESTON J,et al.A semantic matching energy function for learning with multi-relational data[J].Machine Learning,2014,94(2):233-259.
[17]SOCHER R,CHEN D,MANNING C D,et al.Reasoning with neural tensor networks for knowledge base completion[C]//Proc of NIPS 2013.Cambridge,MA:MIT Press,2013:926-934.
[18]BOLLACKER K,EVANS C,PARITOSH P,et al.Freebase:acollaboratively created graph database for structuring human knowledge[C]// Sigmod Conference.2008.
[19]KINGMA D,BA J.Adam:A method for stochastic optimization [J/OL].[2017-01-30].https://arxiv.org/pdf/1412.6980.pdf.
[20]BORDES A,WESTON J,COLLOBERT R,et al.Learningstructured embeddings of knowledge bases[C]//Proc. of AAAI 2011.Menlo Park,CA:AAAI,2011:301-306.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed

一种基于改进向量投影距离的知识图谱表示方法

Knowledge Graph Representation Based on Improved Vector Projection Distance

PDF (PC)

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

Metrics

本文评价

推荐阅读 10

[1]	丁钰, 魏浩, 潘志松, 刘鑫. 网络表示学习算法综述[J]. 计算机科学, 2020, 47(9): 52-59.
[2]	王慧, 乐孜纯, 龚轩, 武玉坤, 左浩. 基于特征分类的链路预测方法综述[J]. 计算机科学, 2020, 47(8): 302-312.
[3]	蒋宗礼, 李苗苗, 张津丽. 基于融合元路径图卷积的异质网络表示学习[J]. 计算机科学, 2020, 47(7): 231-235.
[4]	黄易, 申国伟, 赵文波, 郭春. 一种基于漏洞威胁模式的网络表示学习算法[J]. 计算机科学, 2020, 47(7): 292-298.
[5]	白雪, 努尔布力, 王亚东. 网络安全态势感知研究现状与发展趋势的图谱分析[J]. 计算机科学, 2020, 47(6A): 340-343.
[6]	张志扬, 张凤荔, 陈学勤, 王瑞锦. 基于分层注意力的信息级联预测模型[J]. 计算机科学, 2020, 47(6): 201-209.
[7]	富坤, 仇倩, 赵晓梦, 高金辉. 基于节点演化分阶段优化的事件检测方法[J]. 计算机科学, 2020, 47(5): 96-102.
[8]	袁榕, 宋玉蓉, 孟繁荣. 一种基于加权网络拓扑权重的链路预测方法[J]. 计算机科学, 2020, 47(5): 265-270.
[9]	马扬, 程光权, 梁星星, 李妍, 杨雨灵, 刘忠. 有向加权网络中的改进SDNE算法[J]. 计算机科学, 2020, 47(4): 233-237.
[10]	张虎, 周晶晶, 高海慧, 王鑫. 融合节点结构和内容的网络表示学习方法[J]. 计算机科学, 2020, 47(12): 119-124.
[11]	相颖, 冯钧, 夏珮珮, 陆佳民. 基于Bootstrapping的水利空间关系词提取[J]. 计算机科学, 2020, 47(12): 131-138.
[12]	王慧, 乐孜纯, 龚轩, 左浩, 武玉坤. 基于特征学习的链路预测模型TNTlink[J]. 计算机科学, 2020, 47(12): 245-251.
[13]	吴勇, 王斌君, 翟一鸣, 仝鑫. 共引增强有向网络嵌入研究[J]. 计算机科学, 2020, 47(12): 279-284.
[14]	肖勇, 钱斌, 周密. 基于语义关联的电力计量跨媒体知识图谱构建方法[J]. 计算机科学, 2020, 47(11A): 126-131.
[15]	张春霞, 彭成, 罗妹秋, 牛振东. 数学课程知识图谱构建及其推理[J]. 计算机科学, 2020, 47(11A): 573-578.