Incorporating prior knowledge from counterfactuals into knowledge graph reasoning
Wang, Zikang1,2; Li, Linjing1,2,3; Zeng, Daniel1,2,3; Wu, Xiaofei4
Source PublicationKnowledge-Based Systems
ISSN0950-7051
2021
Volume223Issue:223Pages:TBA
Corresponding AuthorLi, Linjing(linjing.li@ia.ac.cn)
Abstract

Knowledge graph reasoning aims to find the missing links in knowledge graphs and is an important fundamental task. Existing methods mostly reason end-to-end and ignore the prior knowledge in the knowledge graph. In this paper, we attempt to mine prior knowledge from the knowledge graph based on counterfactuals and to use the prior knowledge to enhance the model. Specifically, we begin by constructing counterfactuals to assign a weight for each relation as prior knowledge and then perform reasoning based on both prior knowledge and reinforcement learning. This approach combines the advantages of prior knowledge and neural networks. Experiments on three large datasets show that the prior knowledge extracted from counterfactuals is effective in improving the multi-hop reasoning model. Prior knowledge also has the advantage of being path-length independent, which mitigates the performance degradation in multi-hop reasoning when the reasoning path is excessively long.

Keywordknowledge graph multi-hop reasoning counterfactual
DOI10.1016/j.knosys.2021.107035
WOS KeywordINFERENCE
Indexed BySCI
Language英语
Funding ProjectNational Key Research and Development Program of China[2020AAA0103405] ; National Natural Science Foundation of China[71621002] ; Strategic Priority Research Program of Chinese Academy of Sciences[XDA27030100]
Funding OrganizationNational Key Research and Development Program of China ; National Natural Science Foundation of China ; Strategic Priority Research Program of Chinese Academy of Sciences
WOS Research AreaComputer Science
WOS SubjectComputer Science, Artificial Intelligence
WOS IDWOS:000651271700013
PublisherELSEVIER
Sub direction classification机器学习
Citation statistics
Cited Times:8[WOS]   [WOS Record]     [Related Records in WOS]
Document Type期刊论文
Identifierhttp://ir.ia.ac.cn/handle/173211/44379
Collection多模态人工智能系统全国重点实验室_互联网大数据与信息安全
Affiliation1.State Key Laboratory of Management and Control for Complex Systems, Institute of Automation, Chinese Academy of Sciences
2.Shenzhen Artificial Intelligence and Data Science Institute (Longhua)
3.Beijing Zhongke Wenge Science and Technology Co., Ltd.
4.School of Artificial Intelligence, University of Chinese Academy of Sciences
First Author AffilicationInstitute of Automation, Chinese Academy of Sciences
Recommended Citation
GB/T 7714
Wang, Zikang,Li, Linjing,Zeng, Daniel,et al. Incorporating prior knowledge from counterfactuals into knowledge graph reasoning[J]. Knowledge-Based Systems,2021,223(223):TBA.
APA Wang, Zikang,Li, Linjing,Zeng, Daniel,&Wu, Xiaofei.(2021).Incorporating prior knowledge from counterfactuals into knowledge graph reasoning.Knowledge-Based Systems,223(223),TBA.
MLA Wang, Zikang,et al."Incorporating prior knowledge from counterfactuals into knowledge graph reasoning".Knowledge-Based Systems 223.223(2021):TBA.
Files in This Item: Download All
File Name/Size DocType Version Access License
2021_kbs_counterfact(1353KB)期刊论文作者接受稿开放获取CC BY-NC-SAView
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[Wang, Zikang]'s Articles
[Li, Linjing]'s Articles
[Zeng, Daniel]'s Articles
Baidu academic
Similar articles in Baidu academic
[Wang, Zikang]'s Articles
[Li, Linjing]'s Articles
[Zeng, Daniel]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[Wang, Zikang]'s Articles
[Li, Linjing]'s Articles
[Zeng, Daniel]'s Articles
Terms of Use
No data!
Social Bookmark/Share
File name: 2021_kbs_counterfactualRL.pdf
Format: Adobe PDF
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.