Abstract: When distribution shifts occur between testing and training graph data, out-of-distribution (OOD) samples undermine the performance of graph neural networks (GNNs). To improve adaptive OOD ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results