Abstract: Graph contrastive learning (GCL) leverages semantic consistency as contrastive signals and has shown strong performance in semi-supervised node classification. However, real-world graphs ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results