AutoGCL: Automated Graph Contrastive Learning via Learnable View Generators

Contrastive learning has been widely applied to graph representation learning, where the view generators play a vital role in generating effective contrastive samples. Most of existing methods employ pre-defined generation methods, e.g., node drop or edge perturbation, which usually cannot adapt input data preserve original semantic structures well. To address this issue, we propose novel framework named Automated Graph Learning (AutoGCL) paper. Specifically, AutoGCL employs set learnable orchestrated by an auto augmentation strategy, every generator learns probability distribution graphs conditioned input. While most representative sample, policies introduce adequate variances whole procedure. Furthermore, adopts joint training strategy train generators, encoder, and classifier end-to-end manner, resulting topological heterogeneity yet similarity Extensive experiments on semi-supervised unsupervised transfer demonstrate superiority our over state-of-the-arts learning. In addition, visualization results further confirm that can deliver more compact semantically meaningful samples compared against methods. Our code is available at https://github.com/Somedaywilldo/AutoGCL.