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

Recent years have witnessed great success in handling graph-related taskswith Graph Neural Networks (GNNs). Despite their great academic success,Multi-Layer Perceptrons (MLPs) remain the primary workhorse for practicalindustrial applications. One reason for this academic-industrial gap is theneighborhood-fetching latency incurred by data dependency in GNNs, which makeit hard to deploy for latency-sensitive applications that require fastinference. Conversely, without involving any feature aggregation, MLPs have nodata dependency and infer much faster than GNNs, but their performance is lesscompetitive. Motivated by these complementary strengths and weaknesses, wepropose a Graph Self-Distillation on Neighborhood (GSDN) framework to reducethe gap between GNNs and MLPs. Specifically, the GSDN framework is based purelyon MLPs, where structural information is only implicitly used as prior to guideknowledge self-distillation between the neighborhood and the target,substituting the explicit neighborhood information propagation as in GNNs. As aresult, GSDN enjoys the benefits of graph topology-awareness in training buthas no data dependency in inference. Extensive experiments have shown that theperformance of vanilla MLPs can be greatly improved with self-distillation,e.g., GSDN improves over stand-alone MLPs by 15.54% on average and outperformsthe state-of-the-art GNNs on six datasets. Regarding inference speed, GSDNinfers 75X-89X faster than existing GNNs and 16X-25X faster than otherinference acceleration methods.

Teaching Yourself: Graph Self-Distillation on Neighborhood for Node Classification