RAGAT: Relation Aware Graph Attention Network for Knowledge Graph Completion
Knowledge graph completion (KGC) is the task of predicting missing links based on known triples for knowledge graphs. Several recent works suggest that Graph Neural Networks (GNN) that exploit graph structures achieve promising performance on KGC. These models learn information called messages from...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2021-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9340326/ |
| _version_ | 1818479760880173056 |
|---|---|
| author | Xiyang Liu Huobin Tan Qinghong Chen Guangyan Lin |
| author_facet | Xiyang Liu Huobin Tan Qinghong Chen Guangyan Lin |
| author_sort | Xiyang Liu |
| collection | DOAJ |
| description | Knowledge graph completion (KGC) is the task of predicting missing links based on known triples for knowledge graphs. Several recent works suggest that Graph Neural Networks (GNN) that exploit graph structures achieve promising performance on KGC. These models learn information called messages from neighboring entities and relations and then aggregate messages to update central entity representations. The drawback of existing GNN based models lies in that they tend to treat relations equally and learn fixed network parameters, overlooking the distinction of each relational information. In this work, we propose a <bold>R</bold>elation <bold>A</bold>ware <bold>G</bold>raph <bold>AT</bold>tention network (RAGAT) that constructs separate message functions for different relations, which aims at exploiting the heterogeneous characteristics of knowledge graphs. Specifically, we introduce relation specific parameters to augment the expressive capability of message functions, which enables the model to extract relational information in parameter space. To validate the effect of relation aware mechanism, RAGAT is implemented with a variety of relation aware message functions. Experiments show RAGAT outperforms state-of-the-art link prediction baselines on standard FB15k-237 and WN18RR datasets. |
| first_indexed | 2024-12-10T11:15:01Z |
| format | Article |
| id | doaj.art-d3c6c39329694a8ab90d3967e89a38d8 |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-12-10T11:15:01Z |
| publishDate | 2021-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-d3c6c39329694a8ab90d3967e89a38d82022-12-22T01:51:14ZengIEEEIEEE Access2169-35362021-01-019208402084910.1109/ACCESS.2021.30555299340326RAGAT: Relation Aware Graph Attention Network for Knowledge Graph CompletionXiyang Liu0https://orcid.org/0000-0003-1120-6584Huobin Tan1https://orcid.org/0000-0003-3113-6552Qinghong Chen2https://orcid.org/0000-0001-7471-483XGuangyan Lin3School of Software, Beihang University, Beijing, ChinaSchool of Software, Beihang University, Beijing, ChinaSchool of Software, Beihang University, Beijing, ChinaSchool of Software, Beihang University, Beijing, ChinaKnowledge graph completion (KGC) is the task of predicting missing links based on known triples for knowledge graphs. Several recent works suggest that Graph Neural Networks (GNN) that exploit graph structures achieve promising performance on KGC. These models learn information called messages from neighboring entities and relations and then aggregate messages to update central entity representations. The drawback of existing GNN based models lies in that they tend to treat relations equally and learn fixed network parameters, overlooking the distinction of each relational information. In this work, we propose a <bold>R</bold>elation <bold>A</bold>ware <bold>G</bold>raph <bold>AT</bold>tention network (RAGAT) that constructs separate message functions for different relations, which aims at exploiting the heterogeneous characteristics of knowledge graphs. Specifically, we introduce relation specific parameters to augment the expressive capability of message functions, which enables the model to extract relational information in parameter space. To validate the effect of relation aware mechanism, RAGAT is implemented with a variety of relation aware message functions. Experiments show RAGAT outperforms state-of-the-art link prediction baselines on standard FB15k-237 and WN18RR datasets.https://ieeexplore.ieee.org/document/9340326/Knowledge graph completionknowledge graph embeddinggraph attention networks |
| spellingShingle | Xiyang Liu Huobin Tan Qinghong Chen Guangyan Lin RAGAT: Relation Aware Graph Attention Network for Knowledge Graph Completion IEEE Access Knowledge graph completion knowledge graph embedding graph attention networks |
| title | RAGAT: Relation Aware Graph Attention Network for Knowledge Graph Completion |
| title_full | RAGAT: Relation Aware Graph Attention Network for Knowledge Graph Completion |
| title_fullStr | RAGAT: Relation Aware Graph Attention Network for Knowledge Graph Completion |
| title_full_unstemmed | RAGAT: Relation Aware Graph Attention Network for Knowledge Graph Completion |
| title_short | RAGAT: Relation Aware Graph Attention Network for Knowledge Graph Completion |
| title_sort | ragat relation aware graph attention network for knowledge graph completion |
| topic | Knowledge graph completion knowledge graph embedding graph attention networks |
| url | https://ieeexplore.ieee.org/document/9340326/ |
| work_keys_str_mv | AT xiyangliu ragatrelationawaregraphattentionnetworkforknowledgegraphcompletion AT huobintan ragatrelationawaregraphattentionnetworkforknowledgegraphcompletion AT qinghongchen ragatrelationawaregraphattentionnetworkforknowledgegraphcompletion AT guangyanlin ragatrelationawaregraphattentionnetworkforknowledgegraphcompletion |