Authors

• Kashob Kumar Roy
• Amit Roy
• Amin Ahsan Ali
• M Ashraful Amin
• A K M Mahbubur Rahman

This is a pytorch implementation of our paper "Structure-Aware Hierarchical Graph Pooling using Information Bottleneck" which has been accepted by IJCNN 2021. Check the video presentation of our paper here.

Graph pooling is an essential ingredient of Graph Neural Networks (GNNs) in graph classification and regression tasks. For these tasks, different pooling strategies have been proposed to generate a graph-level representation by downsampling and summarizing nodes' features in a graph. However, most existing pooling methods are unable to capture distinguishable structural information effectively. Besides, they are prone to adversarial attacks. In this work, we propose a novel pooling method named as HIBPool where we leverage the Information Bottleneck IB principle that optimally balances the expressiveness and robustness of a model to learn representations of input data. Furthermore, we introduce a novel structure-aware Discriminative Pooling Readout (DiP-Readout) function to capture the informative local subgraph structures in the graph. Finally, our experimental results show that our model significantly outperforms other state-of-art methods on several graph classification benchmarks and more resilient to feature-perturbation attack than existing pooling methods.

To determine different substructures present in graph data, the input graph is partitioned into communities where intra-community nodes are represented with same color and they are densely connected than inter-community nodes. After aggregating features from neighborhood using Messaging Passing Network (MPN) we apply structure-aware DiP-Readout to obtain distinguishable representations of communities with different substructures while the pooled graph of next layer is created with the edges connecting different communities. Summary representation is obtained by a readout function after the final layer. Optimizing with IB principal ensure minimal redundancy from graph data with sufficient information to predict the class labels.

Table : Comparison with Baselines: ’-’ denotes that results are not publicly available.

If you find our paper or repo useful then please cite our paper:

@article{roy2021structure,
  title={Structure-Aware Hierarchical Graph Pooling using Information Bottleneck},
  author={Roy, Kashob Kumar and Roy, Amit and Rahman, AKM and Amin, M Ashraful and Ali, Amin Ahsan},
  journal={arXiv preprint arXiv:2104.13012},
  year={2021}
}