Abstract: Knowledge distillation (KD), as an effective compression technology, is used to reduce the resource consumption of graph neural networks (GNNs) and facilitate their deployment on ...