SkipNode: On Alleviating Performance Degradation for Deep Graph Convolutional Networks

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lu, Weigang, Zhan, Yibing, Lin, Binbin, Guan, Ziyu, Liu, Liu, Yu, Baosheng, Zhao, Wei, Yang, Yaming, Tao, Dacheng
Format: Preprint
Published: 2021
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916102128271360
author Lu, Weigang
Zhan, Yibing
Lin, Binbin
Guan, Ziyu
Liu, Liu
Yu, Baosheng
Zhao, Wei
Yang, Yaming
Tao, Dacheng
author_facet Lu, Weigang
Zhan, Yibing
Lin, Binbin
Guan, Ziyu
Liu, Liu
Yu, Baosheng
Zhao, Wei
Yang, Yaming
Tao, Dacheng
contents Graph Convolutional Networks (GCNs) suffer from performance degradation when models go deeper. However, earlier works only attributed the performance degeneration to over-smoothing. In this paper, we conduct theoretical and experimental analysis to explore the fundamental causes of performance degradation in deep GCNs: over-smoothing and gradient vanishing have a mutually reinforcing effect that causes the performance to deteriorate more quickly in deep GCNs. On the other hand, existing anti-over-smoothing methods all perform full convolutions up to the model depth. They could not well resist the exponential convergence of over-smoothing due to model depth increasing. In this work, we propose a simple yet effective plug-and-play module, Skipnode, to overcome the performance degradation of deep GCNs. It samples graph nodes in each convolutional layer to skip the convolution operation. In this way, both over-smoothing and gradient vanishing can be effectively suppressed since (1) not all nodes'features propagate through full layers and, (2) the gradient can be directly passed back through ``skipped'' nodes. We provide both theoretical analysis and empirical evaluation to demonstrate the efficacy of Skipnode and its superiority over SOTA baselines.
format Preprint
id arxiv_https___arxiv_org_abs_2112_11628
institution arXiv
publishDate 2021
record_format arxiv
spellingShingle SkipNode: On Alleviating Performance Degradation for Deep Graph Convolutional Networks
Lu, Weigang
Zhan, Yibing
Lin, Binbin
Guan, Ziyu
Liu, Liu
Yu, Baosheng
Zhao, Wei
Yang, Yaming
Tao, Dacheng
Machine Learning
Graph Convolutional Networks (GCNs) suffer from performance degradation when models go deeper. However, earlier works only attributed the performance degeneration to over-smoothing. In this paper, we conduct theoretical and experimental analysis to explore the fundamental causes of performance degradation in deep GCNs: over-smoothing and gradient vanishing have a mutually reinforcing effect that causes the performance to deteriorate more quickly in deep GCNs. On the other hand, existing anti-over-smoothing methods all perform full convolutions up to the model depth. They could not well resist the exponential convergence of over-smoothing due to model depth increasing. In this work, we propose a simple yet effective plug-and-play module, Skipnode, to overcome the performance degradation of deep GCNs. It samples graph nodes in each convolutional layer to skip the convolution operation. In this way, both over-smoothing and gradient vanishing can be effectively suppressed since (1) not all nodes'features propagate through full layers and, (2) the gradient can be directly passed back through ``skipped'' nodes. We provide both theoretical analysis and empirical evaluation to demonstrate the efficacy of Skipnode and its superiority over SOTA baselines.
title SkipNode: On Alleviating Performance Degradation for Deep Graph Convolutional Networks
topic Machine Learning
url https://arxiv.org/abs/2112.11628