Abstract: Graph neural networks (GNNs) are a widely-used class of deep learning (DL) models for learning latent representations of graph-structured data for a variety of node/graph-level prediction ...