Semi-Supervised Text-Attributed Graph Distillation

Proposes a distillation framework to address scalability bottlenecks in representation learning on text-attributed graphs. Leverages semi-supervised learning to utilize both labeled and unlabeled data.