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

Many graph representation learning (GRL) problems are dynamic, with millionsof edges added or removed per second. A fundamental workload in this setting isdynamic link prediction: using a history of graph updates to predict whether agiven pair of vertices will become connected. Recent schemes for linkprediction in such dynamic settings employ Transformers, modeling individualgraph updates as single tokens. In this work, we propose HOT: a model thatenhances this line of works by harnessing higher-order (HO) graph structures;specifically, k-hop neighbors and more general subgraphs containing a givenpair of vertices. Harnessing such HO structures by encoding them into theattention matrix of the underlying Transformer results in higher accuracy oflink prediction outcomes, but at the expense of increased memory pressure. Toalleviate this, we resort to a recent class of schemes that impose hierarchy onthe attention matrix, significantly reducing memory footprint. The final designoffers a sweetspot between high accuracy and low memory utilization. HOToutperforms other dynamic GRL schemes, for example achieving 9%, 7%, and 15%higher accuracy than - respectively - DyGFormer, TGN, and GraphMixer, for theMOOC dataset. Our design can be seamlessly extended towards other dynamic GRLworkloads.

HOT: Higher-Order Dynamic Graph Representation Learning with Efficient Transformers