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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

We consider graph representation learning in a self-supervised manner. Graphneural networks (GNNs) use neighborhood aggregation as a core component thatresults in feature smoothing among nodes in proximity. While successful invarious prediction tasks, such a paradigm falls short of capturing nodes'similarities over a long distance, which proves to be important forhigh-quality learning. To tackle this problem, we strengthen the graph with twoadditional graph views, in which nodes are directly linked to those with themost similar features or local structures. Not restricted by connectivity inthe original graph, the generated views allow the model to enhance itsexpressive power with new and complementary perspectives from which to look atthe relationship between nodes. Following a contrastive learning approach, wepropose a method that aims to maximize the agreement between representationsacross generated views and the original graph. We also propose a channel-levelcontrast approach that greatly reduces computation cost, compared to thecommonly used node level contrast, which requires computation cost quadratic inthe number of nodes. Extensive experiments on seven assortative graphs and fourdisassortative graphs demonstrate the effectiveness of our approach.

Self-Supervised Graph Learning with Proximity-based Views and Channel Contrast