Neighbourhood Transformer: Switchable Attention for Monophily-Aware Graph Learning

Yi Luo, Xu Sun, Guangchun Luo, Aiguo Chen · 2026

Read Neighbourhood Transformer: Switchable Attention for Monophily-Aware Graph Learning by Yi Luo, Xu Sun, Guangchun Luo, Aiguo Chen (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-08980.txt.

Description

Luo and three colleagues propose the Neighbourhood Transformer, a graph neural network built for graphs where linked nodes differ, not for graphs where neighbours resemble each other. Standard graph networks lean on a homophily assumption, the idea that connected nodes are alike, which fails on heterophilic graphs. Drawing on the newly found monophily property of real graphs, the model adds switchable attention that adapts to each neighbourhood. Tests on ten real datasets, five heterophilic and five homophilic, show it beating all leading methods on node classification.

Catalog record

CREATORYi Luo, Xu Sun, Guangchun Luo, Aiguo Chen
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE57466
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.08980

Browse the full library · About the Tower of Babel