Classic GNNs are Strong Baselines: Reassessing GNNs for Node Classification

Our study aims to promote a higher standard of empirical rigor in the field of graph machine learning, encouraging more accurate comparisons and evaluations of model capabilities.