Hypergraph transformer for semi-supervised classification
Hypergraphs play a pivotal role in the modelling of data featuring higher-order relations involving more than two entities. Hypergraph neural networks emerge as a powerful tool for processing hypergraph-structured data, delivering remarkable performance across various tasks, e.g., hypergraph node cl...
| Hauptverfasser: | , , , , , |
|---|---|
| Format: | Conference item |
| Sprache: | English |
| Veröffentlicht: |
IEEE
2024
|
| _version_ | 1826313719491067904 |
|---|---|
| author | Liu, Z Tang, B Ye, Z Dong, X Chen, S Wang, Y |
| author_facet | Liu, Z Tang, B Ye, Z Dong, X Chen, S Wang, Y |
| author_sort | Liu, Z |
| collection | OXFORD |
| description | Hypergraphs play a pivotal role in the modelling of data featuring higher-order relations involving more than two entities. Hypergraph neural networks emerge as a powerful tool for processing hypergraph-structured data, delivering remarkable performance across various tasks, e.g., hypergraph node classification. However, these models struggle to capture global structural information due to their reliance on local message passing. To address this challenge, we propose a novel hypergraph learning framework, HyperGraph Transformer (HyperGT). HyperGT uses a Transformer-based neural network architecture to effectively consider global correlations among all nodes and hyperedges. To incorporate local structural information, HyperGT has two distinct designs: i) a positional encoding based on the hypergraph incidence matrix, offering valuable insights into node-node and hyperedge-hyperedge interactions; and ii) a hypergraph structure regularization in the loss function, capturing connectivities between nodes and hyperedges. Through these designs, HyperGT achieves comprehensive hypergraph representation learning by effectively incorporating global interactions while preserving local connectivity patterns. Extensive experiments conducted on real-world hypergraph node classification tasks showcase that HyperGT consistently outperforms existing methods, establishing new state-of-the-art benchmarks. Ablation studies affirm the effectiveness of the individual designs of our model. |
| first_indexed | 2024-09-25T04:19:23Z |
| format | Conference item |
| id | oxford-uuid:bd07bdad-52db-47c2-8eea-5483fe861c1d |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-09-25T04:19:23Z |
| publishDate | 2024 |
| publisher | IEEE |
| record_format | dspace |
| spelling | oxford-uuid:bd07bdad-52db-47c2-8eea-5483fe861c1d2024-07-30T14:33:04ZHypergraph transformer for semi-supervised classificationConference itemhttp://purl.org/coar/resource_type/c_5794uuid:bd07bdad-52db-47c2-8eea-5483fe861c1dEnglishSymplectic ElementsIEEE2024Liu, ZTang, BYe, ZDong, XChen, SWang, YHypergraphs play a pivotal role in the modelling of data featuring higher-order relations involving more than two entities. Hypergraph neural networks emerge as a powerful tool for processing hypergraph-structured data, delivering remarkable performance across various tasks, e.g., hypergraph node classification. However, these models struggle to capture global structural information due to their reliance on local message passing. To address this challenge, we propose a novel hypergraph learning framework, HyperGraph Transformer (HyperGT). HyperGT uses a Transformer-based neural network architecture to effectively consider global correlations among all nodes and hyperedges. To incorporate local structural information, HyperGT has two distinct designs: i) a positional encoding based on the hypergraph incidence matrix, offering valuable insights into node-node and hyperedge-hyperedge interactions; and ii) a hypergraph structure regularization in the loss function, capturing connectivities between nodes and hyperedges. Through these designs, HyperGT achieves comprehensive hypergraph representation learning by effectively incorporating global interactions while preserving local connectivity patterns. Extensive experiments conducted on real-world hypergraph node classification tasks showcase that HyperGT consistently outperforms existing methods, establishing new state-of-the-art benchmarks. Ablation studies affirm the effectiveness of the individual designs of our model. |
| spellingShingle | Liu, Z Tang, B Ye, Z Dong, X Chen, S Wang, Y Hypergraph transformer for semi-supervised classification |
| title | Hypergraph transformer for semi-supervised classification |
| title_full | Hypergraph transformer for semi-supervised classification |
| title_fullStr | Hypergraph transformer for semi-supervised classification |
| title_full_unstemmed | Hypergraph transformer for semi-supervised classification |
| title_short | Hypergraph transformer for semi-supervised classification |
| title_sort | hypergraph transformer for semi supervised classification |
| work_keys_str_mv | AT liuz hypergraphtransformerforsemisupervisedclassification AT tangb hypergraphtransformerforsemisupervisedclassification AT yez hypergraphtransformerforsemisupervisedclassification AT dongx hypergraphtransformerforsemisupervisedclassification AT chens hypergraphtransformerforsemisupervisedclassification AT wangy hypergraphtransformerforsemisupervisedclassification |