Abstract: Graph neural networks (GNNs) are capable of modeling graph data using various types of nodes and edges, and thus can be widely used in the fields of recommender systems and bioinformatics.