NeurIPS 2022 paper, SubHypergraph Inductive Neural nEtwork

Overview of SHINE: SHINE jointly optimizes the objectives of end-to-end subgraph classification and hypergraph nodes' similarity regularization. SHINE simultaneously learns representations for both nodes and hyperedges using strongly dual attention message passing. The learned representations are aggregated via a subgraph attention layer and used to train a multilayer perceptron for inductive subgraph inferencing.

@inproceedings{luo2022shine,
  title={SHINE: SubHypergraph Inductive Neural nEtwork},
  author={Luo, Yuan},
  booktitle={NeurIPS},
  year={2022}
}

https://arxiv.org/abs/2210.07309

Processed data

Please open an issue or contact [email protected] with any questions.