Training-free Graph Neural Networks and the Power of Labels as Features

arXiv:2404.19288v1 Announce Type: new
Abstract: We propose training-free graph neural networks (TFGNNs), which can be used without training and can also be improved with optional training, for transductive node classification. We first advocate labels as features (LaF), which is an admissible but not explored technique. We show that LaF provably enhances the expressive power of graph neural networks. We design TFGNNs based on this analysis. In the experiments, we confirm that TFGNNs outperform existing GNNs in the training-free setting and …

abstract arxiv classification cs.ai cs.lg features free graph graph neural networks labels networks neural networks node power show stat.ml training type