Relational Structure-Aware Knowledge Graph Representation in Complex Space
Relations in knowledge graphs have rich relational structures and various binary relational patterns. Various relation modelling strategies are proposed for embedding knowledge graphs, but they fail to fully capture both features of relations, rich relational structures and various binary relational...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-06-01
|
| Series: | Mathematics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2227-7390/10/11/1930 |
| _version_ | 1797492583940751360 |
|---|---|
| author | Ke Sun Shuo Yu Ciyuan Peng Yueru Wang Osama Alfarraj Amr Tolba Feng Xia |
| author_facet | Ke Sun Shuo Yu Ciyuan Peng Yueru Wang Osama Alfarraj Amr Tolba Feng Xia |
| author_sort | Ke Sun |
| collection | DOAJ |
| description | Relations in knowledge graphs have rich relational structures and various binary relational patterns. Various relation modelling strategies are proposed for embedding knowledge graphs, but they fail to fully capture both features of relations, rich relational structures and various binary relational patterns. To address the problem of insufficient embedding due to the complexity of the relations, we propose a novel knowledge graph representation model in complex space, namely MARS, to exploit complex relations to embed knowledge graphs. MARS takes the mechanisms of complex numbers and message-passing and then embeds triplets into relation-specific complex hyperplanes. Thus, MARS can well preserve various relation patterns, as well as structural information in knowledge graphs. In addition, we find that the scores generated from the score function approximate a Gaussian distribution. The scores in the tail cannot effectively represent triplets. To address this particular issue and improve the precision of embeddings, we use the standard deviation to limit the dispersion of the score distribution, resulting in more accurate embeddings of triplets. Comprehensive experiments on multiple benchmarks demonstrate that our model significantly outperforms existing state-of-the-art models for link prediction and triple classification. |
| first_indexed | 2024-03-10T01:05:45Z |
| format | Article |
| id | doaj.art-bdd8c148a61e4371a22e34fde5bfeabe |
| institution | Directory Open Access Journal |
| issn | 2227-7390 |
| language | English |
| last_indexed | 2024-03-10T01:05:45Z |
| publishDate | 2022-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Mathematics |
| spelling | doaj.art-bdd8c148a61e4371a22e34fde5bfeabe2023-11-23T14:27:05ZengMDPI AGMathematics2227-73902022-06-011011193010.3390/math10111930Relational Structure-Aware Knowledge Graph Representation in Complex SpaceKe Sun0Shuo Yu1Ciyuan Peng2Yueru Wang3Osama Alfarraj4Amr Tolba5Feng Xia6School of Software, Dalian University of Technology, Dalian 116620, ChinaSchool of Software, Dalian University of Technology, Dalian 116620, ChinaInstitute of Innovation, Science and Sustainability, Federation University Australia, Ballarat, VIC 3353, AustraliaDepartment of Mathematics, National Tsing Hua University, Hsinchu 30013, TaiwanComputer Science Department, Community College, King Saud University, Riyadh 11437, Saudi ArabiaComputer Science Department, Community College, King Saud University, Riyadh 11437, Saudi ArabiaInstitute of Innovation, Science and Sustainability, Federation University Australia, Ballarat, VIC 3353, AustraliaRelations in knowledge graphs have rich relational structures and various binary relational patterns. Various relation modelling strategies are proposed for embedding knowledge graphs, but they fail to fully capture both features of relations, rich relational structures and various binary relational patterns. To address the problem of insufficient embedding due to the complexity of the relations, we propose a novel knowledge graph representation model in complex space, namely MARS, to exploit complex relations to embed knowledge graphs. MARS takes the mechanisms of complex numbers and message-passing and then embeds triplets into relation-specific complex hyperplanes. Thus, MARS can well preserve various relation patterns, as well as structural information in knowledge graphs. In addition, we find that the scores generated from the score function approximate a Gaussian distribution. The scores in the tail cannot effectively represent triplets. To address this particular issue and improve the precision of embeddings, we use the standard deviation to limit the dispersion of the score distribution, resulting in more accurate embeddings of triplets. Comprehensive experiments on multiple benchmarks demonstrate that our model significantly outperforms existing state-of-the-art models for link prediction and triple classification.https://www.mdpi.com/2227-7390/10/11/1930message passingcomplex spaceknowledge representation learninglink predictiontriple classification |
| spellingShingle | Ke Sun Shuo Yu Ciyuan Peng Yueru Wang Osama Alfarraj Amr Tolba Feng Xia Relational Structure-Aware Knowledge Graph Representation in Complex Space Mathematics message passing complex space knowledge representation learning link prediction triple classification |
| title | Relational Structure-Aware Knowledge Graph Representation in Complex Space |
| title_full | Relational Structure-Aware Knowledge Graph Representation in Complex Space |
| title_fullStr | Relational Structure-Aware Knowledge Graph Representation in Complex Space |
| title_full_unstemmed | Relational Structure-Aware Knowledge Graph Representation in Complex Space |
| title_short | Relational Structure-Aware Knowledge Graph Representation in Complex Space |
| title_sort | relational structure aware knowledge graph representation in complex space |
| topic | message passing complex space knowledge representation learning link prediction triple classification |
| url | https://www.mdpi.com/2227-7390/10/11/1930 |
| work_keys_str_mv | AT kesun relationalstructureawareknowledgegraphrepresentationincomplexspace AT shuoyu relationalstructureawareknowledgegraphrepresentationincomplexspace AT ciyuanpeng relationalstructureawareknowledgegraphrepresentationincomplexspace AT yueruwang relationalstructureawareknowledgegraphrepresentationincomplexspace AT osamaalfarraj relationalstructureawareknowledgegraphrepresentationincomplexspace AT amrtolba relationalstructureawareknowledgegraphrepresentationincomplexspace AT fengxia relationalstructureawareknowledgegraphrepresentationincomplexspace |