TR2020-167

Graph Cross Networks with Vertex Infomax Pooling


    •  Li, M., Chen, S., Zhang, Y., "Graph Cross Networks with Vertex Infomax Pooling", Advances in Neural Information Processing Systems (NeurIPS), December 2020.
      BibTeX TR2020-167 PDF
      • @inproceedings{Li2020dec,
      • author = {Li, Maosen and Chen, Siheng and Zhang, Ya},
      • title = {Graph Cross Networks with Vertex Infomax Pooling},
      • booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
      • year = 2020,
      • month = dec,
      • url = {https://www.merl.com/publications/TR2020-167}
      • }
  • Research Areas:

    Artificial Intelligence, Machine Learning, Signal Processing

We propose a novel graph cross network (GXN) to achieve comprehensive feature learning from multiple scales of a graph. Based on trainable hierarchical representations of a graph, GXN enables the interchange of intermediate features across scales to promote information flow. Two key ingredients of GXN include a novel vertex infomax pooling (VIPool), which creates multiscale graphs in a trainable manner, and a novel feature-crossing layer, enabling feature interchange across scales. The proposed VIPool selects the most informative subset of vertices based on the neural estimation of mutual information between vertex features and neighborhood features. The intuition behind is that a vertex is informative when it can maximally reflect its neighboring information. The proposed feature-crossing layer fuses intermediate features between two scales for mutual enhancement by improving information flow and enriching multiscale features at hidden layers. The cross shape of feature-crossing layer distinguishes GXN from many other multiscale architectures. Experimental results show that the proposed GXN improves the classification accuracy by 2.12% and 1.15% on average for graph classification and vertex classification, respectively. Based on the same network, the proposed VIPool consistently outperforms other graph-pooling methods.