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

Graph neural networks (GNN) have been proven to be mature enough for handlinggraph-structured data on node-level graph representation learning tasks.However, the graph pooling technique for learning expressive graph-levelrepresentation is critical yet still challenging. Existing pooling methodseither struggle to capture the local substructure or fail to effectivelyutilize high-order dependency, thus diminishing the expression capability. Inthis paper we propose HAP, a hierarchical graph-level representation learningframework, which is adaptively sensitive to graph structures, i.e., HAPclusters local substructures incorporating with high-order dependencies. HAPutilizes a novel cross-level attention mechanism MOA to naturally focus more onclose neighborhood while effectively capture higher-order dependency that maycontain crucial information. It also learns a global graph content GCont thatextracts the graph pattern properties to make the pre- and post-coarseninggraph content maintain stable, thus providing global guidance in graphcoarsening. This novel innovation also facilitates generalization across graphswith the same form of features. Extensive experiments on fourteen datasets showthat HAP significantly outperforms twelve popular graph pooling methods ongraph classification task with an maximum accuracy improvement of 22.79%, andexceeds the performance of state-of-the-art graph matching and graph similaritylearning algorithms by over 3.5% and 16.7%.

Hierarchical Adaptive Pooling by Capturing High-order Dependency for Graph Representation Learning