MBHAN: Motif-Based Heterogeneous Graph Attention Network
Graph neural networks are graph-based deep learning technologies that have attracted significant attention from researchers because of their powerful performance. Heterogeneous graph-based graph neural networks focus on the heterogeneity of the nodes and links in a graph. This is more effective at p...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-06-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/12/12/5931 |
| _version_ | 1797490346334093312 |
|---|---|
| author | Qian Hu Weiping Lin Minli Tang Jiatao Jiang |
| author_facet | Qian Hu Weiping Lin Minli Tang Jiatao Jiang |
| author_sort | Qian Hu |
| collection | DOAJ |
| description | Graph neural networks are graph-based deep learning technologies that have attracted significant attention from researchers because of their powerful performance. Heterogeneous graph-based graph neural networks focus on the heterogeneity of the nodes and links in a graph. This is more effective at preserving semantic knowledge when representing data interactions in real-world graph structures. Unfortunately, most heterogeneous graph neural networks tend to transform heterogeneous graphs into homogeneous graphs when using meta-paths for representation learning. This paper therefore presents a novel motif-based hierarchical heterogeneous graph attention network algorithm, MBHAN, that addresses this problem by incorporating a hierarchical dual attention mechanism at the node-level and motif-level. Node-level attention aims to learn the importance between a node and its neighboring nodes within its corresponding motif. Motif-level attention is capable of learning the importance of different motifs in the heterogeneous graph. In view of the different vector space features of different types of nodes in heterogeneous graphs, MBHAN also aggregates the features of different types of nodes, so that they can jointly participate in downstream tasks after passing through segregated independent shallow neural networks. MBHAN’s superior network representation learning capability has been validated by extensive experiments on two real-world datasets. |
| first_indexed | 2024-03-10T00:31:40Z |
| format | Article |
| id | doaj.art-95202650abe949c4978fe496b669eede |
| institution | Directory Open Access Journal |
| issn | 2076-3417 |
| language | English |
| last_indexed | 2024-03-10T00:31:40Z |
| publishDate | 2022-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj.art-95202650abe949c4978fe496b669eede2023-11-23T15:24:50ZengMDPI AGApplied Sciences2076-34172022-06-011212593110.3390/app12125931MBHAN: Motif-Based Heterogeneous Graph Attention NetworkQian Hu0Weiping Lin1Minli Tang2Jiatao Jiang3School of Media and Communications, Guizhou Normal University, Guiyang 550000, ChinaSchool of Informatics, Xiamen University, Xiamen 361000, ChinaSchool of Informatics, Xiamen University, Xiamen 361000, ChinaSchool of mathematical Science, Guizhou Normal University, Guiyang 550000, ChinaGraph neural networks are graph-based deep learning technologies that have attracted significant attention from researchers because of their powerful performance. Heterogeneous graph-based graph neural networks focus on the heterogeneity of the nodes and links in a graph. This is more effective at preserving semantic knowledge when representing data interactions in real-world graph structures. Unfortunately, most heterogeneous graph neural networks tend to transform heterogeneous graphs into homogeneous graphs when using meta-paths for representation learning. This paper therefore presents a novel motif-based hierarchical heterogeneous graph attention network algorithm, MBHAN, that addresses this problem by incorporating a hierarchical dual attention mechanism at the node-level and motif-level. Node-level attention aims to learn the importance between a node and its neighboring nodes within its corresponding motif. Motif-level attention is capable of learning the importance of different motifs in the heterogeneous graph. In view of the different vector space features of different types of nodes in heterogeneous graphs, MBHAN also aggregates the features of different types of nodes, so that they can jointly participate in downstream tasks after passing through segregated independent shallow neural networks. MBHAN’s superior network representation learning capability has been validated by extensive experiments on two real-world datasets.https://www.mdpi.com/2076-3417/12/12/5931heterogeneous graphsgraph neural networksrepresentation learningmotifattention mechanism |
| spellingShingle | Qian Hu Weiping Lin Minli Tang Jiatao Jiang MBHAN: Motif-Based Heterogeneous Graph Attention Network Applied Sciences heterogeneous graphs graph neural networks representation learning motif attention mechanism |
| title | MBHAN: Motif-Based Heterogeneous Graph Attention Network |
| title_full | MBHAN: Motif-Based Heterogeneous Graph Attention Network |
| title_fullStr | MBHAN: Motif-Based Heterogeneous Graph Attention Network |
| title_full_unstemmed | MBHAN: Motif-Based Heterogeneous Graph Attention Network |
| title_short | MBHAN: Motif-Based Heterogeneous Graph Attention Network |
| title_sort | mbhan motif based heterogeneous graph attention network |
| topic | heterogeneous graphs graph neural networks representation learning motif attention mechanism |
| url | https://www.mdpi.com/2076-3417/12/12/5931 |
| work_keys_str_mv | AT qianhu mbhanmotifbasedheterogeneousgraphattentionnetwork AT weipinglin mbhanmotifbasedheterogeneousgraphattentionnetwork AT minlitang mbhanmotifbasedheterogeneousgraphattentionnetwork AT jiataojiang mbhanmotifbasedheterogeneousgraphattentionnetwork |