We present a scalable approach for semi-supervised learning on graph-structured data that is based on an efficient variant of convolutional neural networks which operate directly on graphs. We motivate the choice of our convolutional architecture via a localized first-order approximation of spectral graph convolutions. Our model scales linearly in the number of graph edges and learns hidden layer representations that encode both local graph structure and features of nodes.
|Task||Dataset||Model||Metric name||Metric value||Global rank||Compare|
|Node Classification||Citeseer||GCN||Accuracy||70.3%||# 6|
|Node Classification||Cora||GCN||Accuracy||81.5%||# 4|
|Node Classification||NELL||GCN||Accuracy||66.0%||# 1|
|Node Classification||Pubmed||GCN||Accuracy||79.0%||# 5|