Learning on Graphs with Out-of-Distribution Nodes

Yu Song, Donglin Wang

Graph Neural Networks (GNNs) are state-of-the-art models for performing prediction tasks on graphs. While existing GNNs have shown great performance on various tasks related to gra…