Abstract: Graph neural networks (GNNs) have achieved impressive performance when testing and training graph data come from identical distribution. However, existing GNNs lack out-of-distribution ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback