DeepMCGCN: Multi-channel Deep Graph Neural Networks
Abstract Graph neural networks (GNNs) have shown powerful capabilities in modeling and representing graph structural data across various graph learning tasks as an emerging deep learning approach. However, most existing GNNs focus on single-relational graphs and fail to fully utilize the rich and di...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Springer
2024-02-01
|
| Series: | International Journal of Computational Intelligence Systems |
| Subjects: | |
| Online Access: | https://doi.org/10.1007/s44196-024-00432-9 |
| _version_ | 1827326273184071680 |
|---|---|
| author | Lei Meng Zhonglin Ye Yanlin Yang Haixing Zhao |
| author_facet | Lei Meng Zhonglin Ye Yanlin Yang Haixing Zhao |
| author_sort | Lei Meng |
| collection | DOAJ |
| description | Abstract Graph neural networks (GNNs) have shown powerful capabilities in modeling and representing graph structural data across various graph learning tasks as an emerging deep learning approach. However, most existing GNNs focus on single-relational graphs and fail to fully utilize the rich and diverse relational information present in real-world graph data. In addition, deeper GNNs tend to suffer from overfitting and oversmoothing issues, leading to degraded model performance. To deeply excavate the multi-relational features in graph data and strengthen the modeling and representation abilities of GNNs, this paper proposes a multi-channel deep graph convolutional neural network method called DeepMCGCN. It constructs multiple relational subgraphs and adopts multiple GCN channels to learn the characteristics of different relational subgraphs separately. Cross-channel connections are utilized to obtain interactions between different relational subgraphs, which can learn node embeddings richer and more discriminative than single-channel GNNs. Meanwhile, it alleviates overfitting issues of deep models by optimizing convolution functions and adding residual connections between and within channels. The DeepMCGCN method is evaluated on three real-world datasets, and the experimental results show that its node classification performance outperforms that of single-channel GCN and other benchmark models, which improves the modeling and representation capabilities of the model. |
| first_indexed | 2024-03-07T14:42:05Z |
| format | Article |
| id | doaj.art-c8582fe7ebfa4150a6943eb3a14451e7 |
| institution | Directory Open Access Journal |
| issn | 1875-6883 |
| language | English |
| last_indexed | 2024-03-07T14:42:05Z |
| publishDate | 2024-02-01 |
| publisher | Springer |
| record_format | Article |
| series | International Journal of Computational Intelligence Systems |
| spelling | doaj.art-c8582fe7ebfa4150a6943eb3a14451e72024-03-05T20:17:58ZengSpringerInternational Journal of Computational Intelligence Systems1875-68832024-02-0117111310.1007/s44196-024-00432-9DeepMCGCN: Multi-channel Deep Graph Neural NetworksLei Meng0Zhonglin Ye1Yanlin Yang2Haixing Zhao3College of Computer, Qinghai Normal UniversityCollege of Computer, Qinghai Normal UniversityCollege of Computer, Qinghai Normal UniversityCollege of Computer, Qinghai Normal UniversityAbstract Graph neural networks (GNNs) have shown powerful capabilities in modeling and representing graph structural data across various graph learning tasks as an emerging deep learning approach. However, most existing GNNs focus on single-relational graphs and fail to fully utilize the rich and diverse relational information present in real-world graph data. In addition, deeper GNNs tend to suffer from overfitting and oversmoothing issues, leading to degraded model performance. To deeply excavate the multi-relational features in graph data and strengthen the modeling and representation abilities of GNNs, this paper proposes a multi-channel deep graph convolutional neural network method called DeepMCGCN. It constructs multiple relational subgraphs and adopts multiple GCN channels to learn the characteristics of different relational subgraphs separately. Cross-channel connections are utilized to obtain interactions between different relational subgraphs, which can learn node embeddings richer and more discriminative than single-channel GNNs. Meanwhile, it alleviates overfitting issues of deep models by optimizing convolution functions and adding residual connections between and within channels. The DeepMCGCN method is evaluated on three real-world datasets, and the experimental results show that its node classification performance outperforms that of single-channel GCN and other benchmark models, which improves the modeling and representation capabilities of the model.https://doi.org/10.1007/s44196-024-00432-9Deep graph neural networksMulti-relational graphsMulti-channel interactionChannel-level attention mechanism |
| spellingShingle | Lei Meng Zhonglin Ye Yanlin Yang Haixing Zhao DeepMCGCN: Multi-channel Deep Graph Neural Networks International Journal of Computational Intelligence Systems Deep graph neural networks Multi-relational graphs Multi-channel interaction Channel-level attention mechanism |
| title | DeepMCGCN: Multi-channel Deep Graph Neural Networks |
| title_full | DeepMCGCN: Multi-channel Deep Graph Neural Networks |
| title_fullStr | DeepMCGCN: Multi-channel Deep Graph Neural Networks |
| title_full_unstemmed | DeepMCGCN: Multi-channel Deep Graph Neural Networks |
| title_short | DeepMCGCN: Multi-channel Deep Graph Neural Networks |
| title_sort | deepmcgcn multi channel deep graph neural networks |
| topic | Deep graph neural networks Multi-relational graphs Multi-channel interaction Channel-level attention mechanism |
| url | https://doi.org/10.1007/s44196-024-00432-9 |
| work_keys_str_mv | AT leimeng deepmcgcnmultichanneldeepgraphneuralnetworks AT zhonglinye deepmcgcnmultichanneldeepgraphneuralnetworks AT yanlinyang deepmcgcnmultichanneldeepgraphneuralnetworks AT haixingzhao deepmcgcnmultichanneldeepgraphneuralnetworks |