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

Spatial reasoning in text plays a crucial role in various real-worldapplications. Existing approaches for spatial reasoning typically infer spatialrelations from pure text, which overlook the gap between natural language andsymbolic structures. Graph neural networks (GNNs) have showcased exceptionalproficiency in inducing and aggregating symbolic structures. However, classicalGNNs face challenges in handling multi-hop spatial reasoning due to theover-smoothing issue, \textit{i.e.}, the performance decreases substantially asthe number of graph layers increases. To cope with these challenges, we proposea novel \textbf{Dep}th-\textbf{Wi}se \textbf{G}raph \textbf{N}eural\textbf{N}etwork (\textbf{DepWiGNN}). Specifically, we design a novel nodememory scheme and aggregate the information over the depth dimension instead ofthe breadth dimension of the graph, which empowers the ability to collect longdependencies without stacking multiple layers. Experimental results on twochallenging multi-hop spatial reasoning datasets show that DepWiGNN outperformsexisting spatial reasoning methods. The comparisons with the other three GNNsfurther demonstrate its superiority in capturing long dependency in the graph.

DepWiGNN: A Depth-wise Graph Neural Network for Multi-hop Spatial Reasoning in Text