Cargando…
A Block-Based Adaptive Decoupling Framework for Graph Neural Networks
Graph neural networks (GNNs) with feature propagation have demonstrated their power in handling unstructured data. However, feature propagation is also a smooth process that tends to make all node representations similar as the number of propagation increases. To address this problem, we propose a n...
Autores principales: | , , , , |
---|---|
Formato: | Online Artículo Texto |
Lenguaje: | English |
Publicado: |
MDPI
2022
|
Materias: | |
Acceso en línea: | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9497978/ https://www.ncbi.nlm.nih.gov/pubmed/36141076 http://dx.doi.org/10.3390/e24091190 |
_version_ | 1784794641782013952 |
---|---|
author | Shen, Xu Zhang, Yuyang Xie, Yu Wong, Ka-Chun Peng, Chengbin |
author_facet | Shen, Xu Zhang, Yuyang Xie, Yu Wong, Ka-Chun Peng, Chengbin |
author_sort | Shen, Xu |
collection | PubMed |
description | Graph neural networks (GNNs) with feature propagation have demonstrated their power in handling unstructured data. However, feature propagation is also a smooth process that tends to make all node representations similar as the number of propagation increases. To address this problem, we propose a novel Block-Based Adaptive Decoupling (BBAD) Framework to produce effective deep GNNs by utilizing backbone networks. In this framework, each block contains a shallow GNN with feature propagation and transformation decoupled. We also introduce layer regularizations and flexible receptive fields to automatically adjust the propagation depth and to provide different aggregation hops for each node, respectively. We prove that the traditional coupled GNNs are more likely to suffer from over-smoothing when they become deep. We also demonstrate the diversity of outputs from different blocks of our framework. In the experiments, we conduct semi-supervised and fully supervised node classifications on benchmark datasets, and the results verify that our method can not only improve the performance of various backbone networks, but also is superior to existing deep graph neural networks with less parameters. |
format | Online Article Text |
id | pubmed-9497978 |
institution | National Center for Biotechnology Information |
language | English |
publishDate | 2022 |
publisher | MDPI |
record_format | MEDLINE/PubMed |
spelling | pubmed-94979782022-09-23 A Block-Based Adaptive Decoupling Framework for Graph Neural Networks Shen, Xu Zhang, Yuyang Xie, Yu Wong, Ka-Chun Peng, Chengbin Entropy (Basel) Article Graph neural networks (GNNs) with feature propagation have demonstrated their power in handling unstructured data. However, feature propagation is also a smooth process that tends to make all node representations similar as the number of propagation increases. To address this problem, we propose a novel Block-Based Adaptive Decoupling (BBAD) Framework to produce effective deep GNNs by utilizing backbone networks. In this framework, each block contains a shallow GNN with feature propagation and transformation decoupled. We also introduce layer regularizations and flexible receptive fields to automatically adjust the propagation depth and to provide different aggregation hops for each node, respectively. We prove that the traditional coupled GNNs are more likely to suffer from over-smoothing when they become deep. We also demonstrate the diversity of outputs from different blocks of our framework. In the experiments, we conduct semi-supervised and fully supervised node classifications on benchmark datasets, and the results verify that our method can not only improve the performance of various backbone networks, but also is superior to existing deep graph neural networks with less parameters. MDPI 2022-08-25 /pmc/articles/PMC9497978/ /pubmed/36141076 http://dx.doi.org/10.3390/e24091190 Text en © 2022 by the authors. https://creativecommons.org/licenses/by/4.0/Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). |
spellingShingle | Article Shen, Xu Zhang, Yuyang Xie, Yu Wong, Ka-Chun Peng, Chengbin A Block-Based Adaptive Decoupling Framework for Graph Neural Networks |
title | A Block-Based Adaptive Decoupling Framework for Graph Neural Networks |
title_full | A Block-Based Adaptive Decoupling Framework for Graph Neural Networks |
title_fullStr | A Block-Based Adaptive Decoupling Framework for Graph Neural Networks |
title_full_unstemmed | A Block-Based Adaptive Decoupling Framework for Graph Neural Networks |
title_short | A Block-Based Adaptive Decoupling Framework for Graph Neural Networks |
title_sort | block-based adaptive decoupling framework for graph neural networks |
topic | Article |
url | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9497978/ https://www.ncbi.nlm.nih.gov/pubmed/36141076 http://dx.doi.org/10.3390/e24091190 |
work_keys_str_mv | AT shenxu ablockbasedadaptivedecouplingframeworkforgraphneuralnetworks AT zhangyuyang ablockbasedadaptivedecouplingframeworkforgraphneuralnetworks AT xieyu ablockbasedadaptivedecouplingframeworkforgraphneuralnetworks AT wongkachun ablockbasedadaptivedecouplingframeworkforgraphneuralnetworks AT pengchengbin ablockbasedadaptivedecouplingframeworkforgraphneuralnetworks AT shenxu blockbasedadaptivedecouplingframeworkforgraphneuralnetworks AT zhangyuyang blockbasedadaptivedecouplingframeworkforgraphneuralnetworks AT xieyu blockbasedadaptivedecouplingframeworkforgraphneuralnetworks AT wongkachun blockbasedadaptivedecouplingframeworkforgraphneuralnetworks AT pengchengbin blockbasedadaptivedecouplingframeworkforgraphneuralnetworks |