Efficient Graph Neural Network Ensembles for Semi-Supervised Classification

The proposed E2GNN framework effectively assembles multiple graph neural network (GNN) models to achieve improved performance and robustness in semi-supervised node classification tasks.