GANN: Graph Alignment Neural Network for semi-supervised learning

Linxuan Song, Wenxuan Tu, Sihang Zhou, En Zhu

Abstract: Highlights•Efficient mining of features and higher-order neighbor info to reduce oversmoothing.•Three rules improve node representations from features, structure, and predictions.•Comparative experiments use eight datasets, fourteen models, and three label rates.•The experiments provide strong evidence of high-quality node representations.