Tag
This paper introduces a scalable subgraph sampling method using resistance curvature and efficient approximations to reduce training costs for large-scale graph neural networks.