Skip to main navigation Skip to search Skip to main content

Survey on Temporal Knowledge Graph

  • Harbin Institute of Technology

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

Abstract

Knowledge Graph built by people is usually represented as a network with nodes representing entities and edges representing relations between entities. People need to use this form of network architecture to fill in the missing facts in the knowledge graph. Knowledge graph plays an important role in natural language processing. Link prediction task has been studied for a long time as one of the most important tasks in Knowledge Graph Reasoning. Most existing methods focus on static knowledge graph, which cannot get the temporal information in knowledge graph. Now some methods study on temporal knowledge graph, where each fact is associated with a timestamp. The task also becomes more challenging on temporal knowledge graph. In this paper, we introduce some models on static knowledge graph and temporal knowledge graph, and analyze the advantages and disadvantages of each model.

Original languageEnglish
Title of host publicationProceedings - 2021 IEEE 6th International Conference on Data Science in Cyberspace, DSC 2021
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages294-300
Number of pages7
ISBN (Electronic)9781665418157
DOIs
StatePublished - 2021
Externally publishedYes
Event6th IEEE International Conference on Data Science in Cyberspace, DSC 2021 - ShenZhen, China
Duration: 9 Oct 202111 Oct 2021

Publication series

NameProceedings - 2021 IEEE 6th International Conference on Data Science in Cyberspace, DSC 2021

Conference

Conference6th IEEE International Conference on Data Science in Cyberspace, DSC 2021
Country/TerritoryChina
CityShenZhen
Period9/10/2111/10/21

Keywords

  • Deep Learnings
  • Knowledge Graph
  • Knowledge Representation

Fingerprint

Dive into the research topics of 'Survey on Temporal Knowledge Graph'. Together they form a unique fingerprint.

Cite this