English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Contrastive learning has been widely applied to graph representationlearning, where the view generators play a vital role in generating effectivecontrastive samples. Most of the existing contrastive learning methods employpre-defined view generation methods, e.g., node drop or edge perturbation,which usually cannot adapt to input data or preserve the original semanticstructures well. To address this issue, we propose a novel framework namedAutomated Graph Contrastive Learning (AutoGCL) in this paper. Specifically,AutoGCL employs a set of learnable graph view generators orchestrated by anauto augmentation strategy, where every graph view generator learns aprobability distribution of graphs conditioned by the input. While the graphview generators in AutoGCL preserve the most representative structures of theoriginal graph in generation of every contrastive sample, the auto augmentationlearns policies to introduce adequate augmentation variances in the wholecontrastive learning procedure. Furthermore, AutoGCL adopts a joint trainingstrategy to train the learnable view generators, the graph encoder, and theclassifier in an end-to-end manner, resulting in topological heterogeneity yetsemantic similarity in the generation of contrastive samples. Extensiveexperiments on semi-supervised learning, unsupervised learning, and transferlearning demonstrate the superiority of our AutoGCL framework over thestate-of-the-arts in graph contrastive learning. In addition, the visualizationresults further confirm that the learnable view generators can deliver morecompact and semantically meaningful contrastive samples compared against theexisting view generation methods.

AutoGCL: Automated Graph Contrastive Learning via Learnable View Generators