Teaching MLP More Graph Information: A Three-stage Multitask Knowledge Distillation Framework