CN-Motifs Perceptive Graph Neural Networks

Graph neural networks (GNNs) have been the dominant approaches for graph representation learning. However, most GNNs are applied to homophily graphs and perform poorly on heterophily graphs. Meanwhile, these GNNs fail to directly capture long-range dependencies and complex interactions between 1-hop...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Fan Zhang, Tian-Ming Bu
Formato:	article
Lenguaje:	EN
Publicado:	IEEE 2021
Materias:	Common neighbors graph neural networks heterophily graph motifs structural patterns Electrical engineering. Electronics. Nuclear engineering TK1-9971
Acceso en línea:	https://doaj.org/article/270e379a7473422bb382ef7305a6eb05
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:270e379a7473422bb382ef7305a6eb05
record_format	dspace
spelling	oai:doaj.org-article:270e379a7473422bb382ef7305a6eb052021-11-17T00:01:04ZCN-Motifs Perceptive Graph Neural Networks2169-353610.1109/ACCESS.2021.3126417https://doaj.org/article/270e379a7473422bb382ef7305a6eb052021-01-01T00:00:00Zhttps://ieeexplore.ieee.org/document/9606682/https://doaj.org/toc/2169-3536Graph neural networks (GNNs) have been the dominant approaches for graph representation learning. However, most GNNs are applied to homophily graphs and perform poorly on heterophily graphs. Meanwhile, these GNNs fail to directly capture long-range dependencies and complex interactions between 1-hop neighbors when generating node representations by iteratively aggregating directly connected neighbors. In addition, structural patterns, such as motifs which have been established as building blocks for graph structure, contain rich topological and semantical information and are worth studying further. In this paper, we introduce the common-neighbors based motifs, which we called CN-motifs, to generalize and enrich the definition of structural patterns. We group the 1-hop neighbors and construct a high-order graph according to CN-motifs, and propose CN-motifs Perceptive Graph Neural Networks (CNMPGNN), a novel framework which can effectively resolve problems mentioned above. Notably, by making full use of structural patterns, our model achieves the state-of-the-art results on several homophily and heterophily datasets.Fan ZhangTian-Ming BuIEEEarticleCommon neighborsgraph neural networksheterophily graphmotifsstructural patternsElectrical engineering. Electronics. Nuclear engineeringTK1-9971ENIEEE Access, Vol 9, Pp 151285-151293 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Common neighbors graph neural networks heterophily graph motifs structural patterns Electrical engineering. Electronics. Nuclear engineering TK1-9971
spellingShingle	Common neighbors graph neural networks heterophily graph motifs structural patterns Electrical engineering. Electronics. Nuclear engineering TK1-9971 Fan Zhang Tian-Ming Bu CN-Motifs Perceptive Graph Neural Networks
description	Graph neural networks (GNNs) have been the dominant approaches for graph representation learning. However, most GNNs are applied to homophily graphs and perform poorly on heterophily graphs. Meanwhile, these GNNs fail to directly capture long-range dependencies and complex interactions between 1-hop neighbors when generating node representations by iteratively aggregating directly connected neighbors. In addition, structural patterns, such as motifs which have been established as building blocks for graph structure, contain rich topological and semantical information and are worth studying further. In this paper, we introduce the common-neighbors based motifs, which we called CN-motifs, to generalize and enrich the definition of structural patterns. We group the 1-hop neighbors and construct a high-order graph according to CN-motifs, and propose CN-motifs Perceptive Graph Neural Networks (CNMPGNN), a novel framework which can effectively resolve problems mentioned above. Notably, by making full use of structural patterns, our model achieves the state-of-the-art results on several homophily and heterophily datasets.
format	article
author	Fan Zhang Tian-Ming Bu
author_facet	Fan Zhang Tian-Ming Bu
author_sort	Fan Zhang
title	CN-Motifs Perceptive Graph Neural Networks
title_short	CN-Motifs Perceptive Graph Neural Networks
title_full	CN-Motifs Perceptive Graph Neural Networks
title_fullStr	CN-Motifs Perceptive Graph Neural Networks
title_full_unstemmed	CN-Motifs Perceptive Graph Neural Networks
title_sort	cn-motifs perceptive graph neural networks
publisher	IEEE
publishDate	2021
url	https://doaj.org/article/270e379a7473422bb382ef7305a6eb05
work_keys_str_mv	AT fanzhang cnmotifsperceptivegraphneuralnetworks AT tianmingbu cnmotifsperceptivegraphneuralnetworks
_version_	1718426036666892288

CN-Motifs Perceptive Graph Neural Networks

Ejemplares similares