Skip to main navigation Skip to search Skip to main content

Compressing Knowledge Graph Embedding with Relational Graph Auto-encoder

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Knowledge graphs (KGs) are extremely useful resources for varieties of applications. However, with the large and steadily growing sizes of modern KGs, knowledge graph embeddings (KGE), which represent entities and relations in KGs into 32-bit floating-point vectors, become more and more expensive in terms of memory. To this end, in this paper, we propose a general framework to compress the embeddings from real-valued vectors to binary ones while preserving the inherent information of KGs. Specifically, the proposed framework utilizes relational graph auto-encoders as well as the Gumbel-Softmax trick to obtain the compressed representations. Our framework can be applied to a number of existing KGE models. Particularly, we extend state-of-the-art models TransE, DistMult, and ConvE in this paper. Finally, extensive experiments show that the proposed method successfully reduces the memory size of the embeddings by 92% while only leading to a loss of no more than 5% in the knowledge graph completion task.

Original languageEnglish
Title of host publicationICEIEC 2020 - Proceedings of 2020 IEEE 10th International Conference on Electronics Information and Emergency Communication
EditorsWenzheng Li, Xuefei Zhang
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages366-370
Number of pages5
ISBN (Electronic)9781728163123
DOIs
StatePublished - Jul 2020
Externally publishedYes
Event10th IEEE International Conference on Electronics Information and Emergency Communication, ICEIEC 2020 - Beijing, China
Duration: 17 Jul 202019 Jul 2020

Publication series

NameICEIEC 2020 - Proceedings of 2020 IEEE 10th International Conference on Electronics Information and Emergency Communication

Conference

Conference10th IEEE International Conference on Electronics Information and Emergency Communication, ICEIEC 2020
Country/TerritoryChina
CityBeijing
Period17/07/2019/07/20

Keywords

  • Compression
  • Graph autoencoders
  • Knowledge graph completion
  • Knowledge graph embedding

Fingerprint

Dive into the research topics of 'Compressing Knowledge Graph Embedding with Relational Graph Auto-encoder'. Together they form a unique fingerprint.

Cite this