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

While contrastive self-supervised learning has become the de-facto learningparadigm for graph neural networks, the pursuit of high task accuracy requiresa large hidden dimensionality to learn informative and discriminativefull-precision representations, raising concerns about computation, memoryfootprint, and energy consumption burden (largely overlooked) for real-worldapplications. This paper explores a promising direction for graph contrastivelearning (GCL) with spiking neural networks (SNNs), which leverage sparse andbinary characteristics to learn more biologically plausible and compactrepresentations. We propose SpikeGCL, a novel GCL framework to learn binarized1-bit representations for graphs, making balanced trade-offs between efficiencyand performance. We provide theoretical guarantees to demonstrate that SpikeGCLhas comparable expressiveness with its full-precision counterparts.Experimental results demonstrate that, with nearly 32x representation storagecompression, SpikeGCL is either comparable to or outperforms many fancystate-of-the-art supervised and self-supervised methods across several graphbenchmarks.

A Graph is Worth 1-bit Spikes: When Graph Contrastive Learning Meets Spiking Neural Networks