Self-supervised graph representations with generative adversarial learning

Xuecheng Sun, Zonghui Wang, Zheming Lu, Ziqian Lu

Abstract: Highlights•Generating attributed subgraphs in a latent space for graph representation learning.•Capturing both the local and global information of the subgraph samples.•Technical tricks to make the discrete generation process learnable.•Learning generalized representations for node classification and link prediction.