Explicit and Implicit Feature Contrastive Learning Model for Knowledge Graph Link Prediction

Knowledge graph link prediction is crucial for constructing triples in knowledge graphs, which aim to infer whether there is a relation between the entities. Recently, graph neural networks and contrastive learning have demonstrated superior performance compared with traditional translation-based mo...

Full description

Saved in:
Bibliographic Details
Main Authors: Xu Yuan, Weihe Wang, Buyun Gao, Liang Zhao, Ruixin Ma, Feng Ding
Format: Article
Language:English
Published: MDPI AG 2024-11-01
Series:Sensors
Subjects:
Online Access:https://www.mdpi.com/1424-8220/24/22/7353
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1846152398020542464
author Xu Yuan
Weihe Wang
Buyun Gao
Liang Zhao
Ruixin Ma
Feng Ding
author_facet Xu Yuan
Weihe Wang
Buyun Gao
Liang Zhao
Ruixin Ma
Feng Ding
author_sort Xu Yuan
collection DOAJ
description Knowledge graph link prediction is crucial for constructing triples in knowledge graphs, which aim to infer whether there is a relation between the entities. Recently, graph neural networks and contrastive learning have demonstrated superior performance compared with traditional translation-based models; they successfully extracted common features through explicit linking between entities. However, the implicit associations between entities without a linking relationship are ignored, which impedes the model from capturing distant but semantically rich entities. In addition, directly applying contrastive learning based on random node dropout to link prediction tasks, or limiting it to triplet-level, leads to constrained model performance. To address these challenges, we design an implicit feature extraction module that utilizes the clustering characteristics of latent vector space to find entities with potential associations and enrich entity representations by mining similar semantic features from the conceptual level. Meanwhile, the subgraph mechanism is introduced to preserve the structural information of explicitly connected entities. Implicit semantic features and explicit structural features serve as complementary information to provide high-quality self-supervised signals. Experiments are conducted on three benchmark knowledge graph datasets. The results validate that our model outperforms the state-of-the-art baselines in link prediction tasks.
format Article
id doaj-art-200c2be26bed44bf98ca6eb6d814ebfe
institution Kabale University
issn 1424-8220
language English
publishDate 2024-11-01
publisher MDPI AG
record_format Article
series Sensors
spelling doaj-art-200c2be26bed44bf98ca6eb6d814ebfe2024-11-26T18:21:40ZengMDPI AGSensors1424-82202024-11-012422735310.3390/s24227353Explicit and Implicit Feature Contrastive Learning Model for Knowledge Graph Link PredictionXu Yuan0Weihe Wang1Buyun Gao2Liang Zhao3Ruixin Ma4Feng Ding5School of Software Technology, Dalian University of Technology, Dalian 116024, ChinaSchool of Software Technology, Dalian University of Technology, Dalian 116024, ChinaSchool of Software Technology, Dalian University of Technology, Dalian 116024, ChinaSchool of Software Technology, Dalian University of Technology, Dalian 116024, ChinaSchool of Software Technology, Dalian University of Technology, Dalian 116024, ChinaSchool of Software Technology, Dalian University of Technology, Dalian 116024, ChinaKnowledge graph link prediction is crucial for constructing triples in knowledge graphs, which aim to infer whether there is a relation between the entities. Recently, graph neural networks and contrastive learning have demonstrated superior performance compared with traditional translation-based models; they successfully extracted common features through explicit linking between entities. However, the implicit associations between entities without a linking relationship are ignored, which impedes the model from capturing distant but semantically rich entities. In addition, directly applying contrastive learning based on random node dropout to link prediction tasks, or limiting it to triplet-level, leads to constrained model performance. To address these challenges, we design an implicit feature extraction module that utilizes the clustering characteristics of latent vector space to find entities with potential associations and enrich entity representations by mining similar semantic features from the conceptual level. Meanwhile, the subgraph mechanism is introduced to preserve the structural information of explicitly connected entities. Implicit semantic features and explicit structural features serve as complementary information to provide high-quality self-supervised signals. Experiments are conducted on three benchmark knowledge graph datasets. The results validate that our model outperforms the state-of-the-art baselines in link prediction tasks.https://www.mdpi.com/1424-8220/24/22/7353knowledge graphlink predictioncontrastive learningimplicit semantic feature
spellingShingle Xu Yuan
Weihe Wang
Buyun Gao
Liang Zhao
Ruixin Ma
Feng Ding
Explicit and Implicit Feature Contrastive Learning Model for Knowledge Graph Link Prediction
Sensors
knowledge graph
link prediction
contrastive learning
implicit semantic feature
title Explicit and Implicit Feature Contrastive Learning Model for Knowledge Graph Link Prediction
title_full Explicit and Implicit Feature Contrastive Learning Model for Knowledge Graph Link Prediction
title_fullStr Explicit and Implicit Feature Contrastive Learning Model for Knowledge Graph Link Prediction
title_full_unstemmed Explicit and Implicit Feature Contrastive Learning Model for Knowledge Graph Link Prediction
title_short Explicit and Implicit Feature Contrastive Learning Model for Knowledge Graph Link Prediction
title_sort explicit and implicit feature contrastive learning model for knowledge graph link prediction
topic knowledge graph
link prediction
contrastive learning
implicit semantic feature
url https://www.mdpi.com/1424-8220/24/22/7353
work_keys_str_mv AT xuyuan explicitandimplicitfeaturecontrastivelearningmodelforknowledgegraphlinkprediction
AT weihewang explicitandimplicitfeaturecontrastivelearningmodelforknowledgegraphlinkprediction
AT buyungao explicitandimplicitfeaturecontrastivelearningmodelforknowledgegraphlinkprediction
AT liangzhao explicitandimplicitfeaturecontrastivelearningmodelforknowledgegraphlinkprediction
AT ruixinma explicitandimplicitfeaturecontrastivelearningmodelforknowledgegraphlinkprediction
AT fengding explicitandimplicitfeaturecontrastivelearningmodelforknowledgegraphlinkprediction