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

When handling streaming graphs, existing graph representation learning modelsencounter a catastrophic forgetting problem, where previously learned knowledgeof these models is easily overwritten when learning with newly incoming graphs.In response, Continual Graph Learning emerges as a novel paradigm enablinggraph representation learning from static to streaming graphs. Our prior work,CaT is a replay-based framework with a balanced continual learning procedure,which designs a small yet effective memory bank for replaying data bycondensing incoming graphs. Although the CaT alleviates the catastrophicforgetting problem, there exist three issues: (1) The graph condensationalgorithm derived in CaT only focuses on labelled nodes while neglectingabundant information carried by unlabelled nodes; (2) The continual trainingscheme of the CaT overemphasises on the previously learned knowledge, limitingthe model capacity to learn from newly added memories; (3) Both thecondensation process and replaying process of the CaT are time-consuming. Inthis paper, we propose a psudo-label guided memory bank (PUMA) CGL framework,extending from the CaT to enhance its efficiency and effectiveness byovercoming the above-mentioned weaknesses and limits. To fully exploit theinformation in a graph, PUMA expands the coverage of nodes during graphcondensation with both labelled and unlabelled nodes. Furthermore, atraining-from-scratch strategy is proposed to upgrade the previous continuallearning scheme for a balanced training between the historical and the newgraphs. Besides, PUMA uses a one-time prorogation and wide graph encoders toaccelerate the graph condensation and the graph encoding process in thetraining stage to improve the efficiency of the whole framework. Extensiveexperiments on four datasets demonstrate the state-of-the-art performance andefficiency over existing methods.

PUMA: Efficient Continual Graph Learning with Graph Condensation