Positionally restricted masked knowledge graph completion via multi-head mutual attention

Knowledge graph completion aims to enhance the completeness of knowledge graphs by predicting missing links. Link prediction is a common approach for this task, but existing methods, particularly those based on similarity computation, are often computationally expensive, especially for large models....

Full description

Saved in:
Bibliographic Details
Main Authors: Qiang Yu, Liang Bao, Peng Nie, Lei Zuo
Format: Article
Language:English
Published: KeAi Communications Co., Ltd. 2025-05-01
Series:Journal of Information and Intelligence
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2949715925000095
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849472851210403840
author Qiang Yu
Liang Bao
Peng Nie
Lei Zuo
author_facet Qiang Yu
Liang Bao
Peng Nie
Lei Zuo
author_sort Qiang Yu
collection DOAJ
description Knowledge graph completion aims to enhance the completeness of knowledge graphs by predicting missing links. Link prediction is a common approach for this task, but existing methods, particularly those based on similarity computation, are often computationally expensive, especially for large models. To address this, we propose a novel method, positionally restricted masked knowledge graph completion (PR-MKGC), which reduces inference time by leveraging masked prediction and relying solely on structural information from the knowledge graph, without using textual data. We introduce a multi-head mutual attention mechanism that aggregates neighbor information more effectively, improving the model's ability to predict missing links. Experimental results demonstrate that PR-MKGC outperforms existing models in terms of both predictive performance and inference time on the FB15K-237 dataset.
format Article
id doaj-art-bb6221b07c3d474c8f7d5f0786bad17d
institution Kabale University
issn 2949-7159
language English
publishDate 2025-05-01
publisher KeAi Communications Co., Ltd.
record_format Article
series Journal of Information and Intelligence
spelling doaj-art-bb6221b07c3d474c8f7d5f0786bad17d2025-08-20T03:24:22ZengKeAi Communications Co., Ltd.Journal of Information and Intelligence2949-71592025-05-013321022210.1016/j.jiixd.2025.02.006Positionally restricted masked knowledge graph completion via multi-head mutual attentionQiang Yu0Liang Bao1Peng Nie2Lei Zuo3Hangzhou Institute of Technology, Xidian University, Hangzhou 311231, China; School of Computer Science and Technology, Xidian University, Xi'an 710071, ChinaSchool of Computer Science and Technology, Xidian University, Xi'an 710071, ChinaHangzhou Institute of Technology, Xidian University, Hangzhou 311231, China; Corresponding author.National Lab of Radar Signal Processing, Xidian University, Xi'an 710071, China; Corresponding author.Knowledge graph completion aims to enhance the completeness of knowledge graphs by predicting missing links. Link prediction is a common approach for this task, but existing methods, particularly those based on similarity computation, are often computationally expensive, especially for large models. To address this, we propose a novel method, positionally restricted masked knowledge graph completion (PR-MKGC), which reduces inference time by leveraging masked prediction and relying solely on structural information from the knowledge graph, without using textual data. We introduce a multi-head mutual attention mechanism that aggregates neighbor information more effectively, improving the model's ability to predict missing links. Experimental results demonstrate that PR-MKGC outperforms existing models in terms of both predictive performance and inference time on the FB15K-237 dataset.http://www.sciencedirect.com/science/article/pii/S2949715925000095Knowledge graphLink predictionAttention mechanismInformation aggregation
spellingShingle Qiang Yu
Liang Bao
Peng Nie
Lei Zuo
Positionally restricted masked knowledge graph completion via multi-head mutual attention
Journal of Information and Intelligence
Knowledge graph
Link prediction
Attention mechanism
Information aggregation
title Positionally restricted masked knowledge graph completion via multi-head mutual attention
title_full Positionally restricted masked knowledge graph completion via multi-head mutual attention
title_fullStr Positionally restricted masked knowledge graph completion via multi-head mutual attention
title_full_unstemmed Positionally restricted masked knowledge graph completion via multi-head mutual attention
title_short Positionally restricted masked knowledge graph completion via multi-head mutual attention
title_sort positionally restricted masked knowledge graph completion via multi head mutual attention
topic Knowledge graph
Link prediction
Attention mechanism
Information aggregation
url http://www.sciencedirect.com/science/article/pii/S2949715925000095
work_keys_str_mv AT qiangyu positionallyrestrictedmaskedknowledgegraphcompletionviamultiheadmutualattention
AT liangbao positionallyrestrictedmaskedknowledgegraphcompletionviamultiheadmutualattention
AT pengnie positionallyrestrictedmaskedknowledgegraphcompletionviamultiheadmutualattention
AT leizuo positionallyrestrictedmaskedknowledgegraphcompletionviamultiheadmutualattention