ABSTRACT
In recent years, different web knowledge graphs, both free and commercial, have been created. Knowledge graphs use relations between entities to describe facts in the world. We engage in embedding a large scale knowledge graph into a continuous vector space. TransE, TransH, TransR and TransD are promising methods proposed in recent years and achieved state-of-the-art predictive performance. In this paper, we discuss that graph structures should be considered in embedding and propose to embed substructures called "one-relation-circle" (ORC) to further improve the performance of the above methods as they are unable to encode ORC substructures. Some complex models are capable of handling ORC structures but sacrifice efficiency in the process. To make a good trade-off between the model capacity and efficiency, we propose a method to decompose ORC substructures by using two vectors to represent the entity as a head or tail entity with the same relation. In this way, we can encode the ORC structure properly when apply it to TransH, TransR and TransD with almost the same model complexity of themselves. We conduct experiments on link prediction with benchmark dataset WordNet. Our experiments show that applying our method improves the results compared with the corresponding original results of TransH, TransR and TransD.
- J. W. A. Bordes, X. Glorot and Y. Bengio. A semantic matching energy function for learning with multi-relational data. Machine Learning, 2013. Google ScholarDigital Library
- R. C. A. Bordes, J.Weston and Y. Bengio. Learning structured embeddings of knowledge bases. AAAI, 2011. Google ScholarDigital Library
- N. U. Alberto Garct'Asa-Durt'an, Antoine Bordes and Y. Grandvalet. Combining two and three-way embedding models for link prediction in knowledge bases. Journal of Artificial Intelligence Research, 2016. Google ScholarDigital Library
- B. K. B. S. E. R. H. J. Andrew Carlson, Justin Betteridge and T. M. Mitchell1. Toward an architecture for never-ending language learning. AAAI, 2010. Google ScholarDigital Library
- R. C. Antoine Bordes, JasonWeston and Y. Bengio. Learning structured embeddings of knowledge bases. AAAI, 2011. Google ScholarDigital Library
- U. N. G.-D. A. W. J. Bordes, Antoine and O. Yakhnenko. Translating embeddings for modeling multi-relational data. NIPS, 2013b. Google ScholarDigital Library
- S. B. Denis Krompa and V. Tresp. Type-constrained representation learning in knowledge graphs. ISWC, 2015. Google ScholarDigital Library
- G. K. F. M. Suchanek and G. Weikum. Yago: a core of semantic knowledge. WWW, 2007. Google ScholarDigital Library
- L. X. K. L. Guoliang Ji, Shizhu He and J. Zhao. Knowledge graph embedding via dynamic mapping matrix. ACL, 2015.Google Scholar
- M. H. Han Xiao and X. Zhu. Ssp: Semantic space projection for knowledge graph embedding with text descriptions. AAAI, 2017.Google Scholar
- P. P. T. S. K. Bollacker, C. Evans and J. Taylor. Freebase: A collaboratively created graph database for structuring human knowledge. SIGMOD, 2008. Google ScholarDigital Library
- L. R. M. Nickel and P. Tomaso. Holographic embeddings of knowledge graphs. AAAI, 2016. Google ScholarDigital Library
- V. T. M. Nickel and H.-P. Kriegel. A three-way model for collective learning on multi- relational data. ICML, 2011.Google Scholar
- B. R. F. C. G. D. M. K. Miller, G.A. Introduction to wordnet: An on-line lexical database. International journal of lexicography, 1990. Google ScholarCross Ref
- G. A. Miller. Wordnet: A lexical database for english. Communications of the ACM Vol. 38, 1995. Google ScholarDigital Library
- F. P. Ni Lao, Amarnag Subramanya and W. Cohen. Reading the web with learned syntactic-semantic inference rules. EMNLP, 2012. Google ScholarDigital Library
- A. B. G. O. R. Jenatton, N. Le Roux. A latent factor model for highly multi-relational data. AISTATS, 2010.Google Scholar
- H. Robbins and Monro. A stochastic approximation method. Annals of Mathematical Statistics, 1951.Google ScholarCross Ref
- Z. L. Ruobing Xie and M. Sun. Representation learning of knowledge graphs with hierarchical types. IJCAI, 2016. Google ScholarDigital Library
- B. Shi and T. Weninger. Proje:embedding projection for knowledge graph completion. AAAI, 2017.Google Scholar
- G. J. Shizhu He, Kang Liu and J. Zhao. Learning to represent knowledge graphs with gaussian embedding. CIKM, 2015. Google ScholarDigital Library
- C. D. M. C. D. Socher, Richard and A. Y. Ng. Reasoning with neural tensor networks for knowledge base completion. NIPS, 2013. Google ScholarDigital Library
- M. Xiao, H.; Huang and X. Zhu. Transg: A generative model for knowledge graph embedding. ACL, 2016.Google ScholarCross Ref
- Z. J. J. L. H. Xie, R.; Liu and M. Sun. Representation learning of knowledge graphs with entity descriptions. AAAI, 2016. Google ScholarDigital Library
- H. L. X. J. Y. Jia, Y. Wang and X. Cheng. Locally adaptive translation for knowledge graph embedding. AAAI, 2016. Google ScholarDigital Library
- M. S. Y. L. Y. Lin, Z. Liu and X. Zhu. Learning entity and relation embeddings for knowledge graph completion. AAAI, 2015. Google ScholarDigital Library
- Z. L. Y. Lin and M. Sun. Modeling relation paths for representation learning of knowledge bases. EMNLP, 2015. Google ScholarCross Ref
- M. S. Yankai Lin, Zhiyuan Liu. Knowledge representation learning with entities, attributes and relations. IJCAI, 2016. Google ScholarDigital Library
- J. F. Z. Wang, J. Zhang and Z. Chen. Knowledge graph embedding by translating on hyperplanes. AAAI, 2014. Google ScholarDigital Library
Index Terms
- Knowledge Graph Embedding with Diversity of Structures
Recommendations
ProjR: Embedding Structure Diversity for Knowledge Graph Completion
Natural Language Processing and Chinese ComputingAbstractKnowledge graph completion aims to find new true links between entities. In this paper, we consider an approach to embed a knowledge graph into a continuous vector space. Embedding methods, such as TransE, TransR and ProjE, are proposed in recent ...
Re-evaluating Embedding-Based Knowledge Graph Completion Methods
CIKM '18: Proceedings of the 27th ACM International Conference on Information and Knowledge ManagementIncompleteness of large knowledge graphs (KG) has motivated many researchers to propose methods to automatically find missing edges in KGs. A promising approach for KG completion (link prediction) is embedding a KG into a continuous vector space. There ...
SSKGE: a time-saving knowledge graph embedding framework based on structure enhancement and semantic guidance
AbstractIn knowledge graph embedding, an attempt is made to embed the objective facts and relationships expressed in the form of triplets into multidimensional vector space, facilitating various applications, such as link prediction and question ...
Comments