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

Temporal graph is an abstraction for modeling dynamic systems that consist ofevolving interaction elements. In this paper, we aim to solve an important yetneglected problem -- how to learn information from high-order neighbors intemporal graphs? -- to enhance the informativeness and discriminativeness forthe learned node representations. We argue that when learning high-orderinformation from temporal graphs, we encounter two challenges, i.e.,computational inefficiency and over-smoothing, that cannot be solved byconventional techniques applied on static graphs. To remedy these deficiencies,we propose a temporal propagation-based graph neural network, namely TPGNN. Tobe specific, the model consists of two distinct components, i.e., propagatorand node-wise encoder. The propagator is leveraged to propagate messages fromthe anchor node to its temporal neighbors within $k$-hop, and thensimultaneously update the state of neighborhoods, which enables efficientcomputation, especially for a deep model. In addition, to preventover-smoothing, the model compels the messages from $n$-hop neighbors to updatethe $n$-hop memory vector preserved on the anchor. The node-wise encoder adoptstransformer architecture to learn node representations by explicitly learningthe importance of memory vectors preserved on the node itself, that is,implicitly modeling the importance of messages from neighbors at differentlayers, thus mitigating the over-smoothing. Since the encoding process will notquery temporal neighbors, we can dramatically save time consumption ininference. Extensive experiments on temporal link prediction and nodeclassification demonstrate the superiority of TPGNN over state-of-the-artbaselines in efficiency and robustness.

TPGNN: Learning High-order Information in Dynamic Graphs via Temporal Propagation