A class-aware representation refinement framework for graph classification

Graph Neural Networks (GNNs) are widely used for graph representation learning. Despite its prevalence, GNN suffers from two drawbacks in the graph classification task, the neglect of graph-level relationships, and the generalization issue. Each graph is treated separately in GNN message passing/gra...

Full description

Saved in:
Bibliographic Details
Main Authors: Xu, Jiaxing, Ni, Jinjie, Ke, Yiping
Other Authors: College of Computing and Data Science
Format: Article
Language:English
Published: 2024
Subjects:
Online Access:https://hdl.handle.net/10356/180546
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
id sg-ntu-dr.10356-180546
record_format dspace
spelling sg-ntu-dr.10356-1805462024-10-14T08:48:26Z A class-aware representation refinement framework for graph classification Xu, Jiaxing Ni, Jinjie Ke, Yiping College of Computing and Data Science School of Computer Science and Engineering Computer and Information Science Graph neural network Representation learning Graph classification Graph Neural Networks (GNNs) are widely used for graph representation learning. Despite its prevalence, GNN suffers from two drawbacks in the graph classification task, the neglect of graph-level relationships, and the generalization issue. Each graph is treated separately in GNN message passing/graph pooling, and existing methods to address overfitting operate on each individual graph. This makes the graph representations learnt less effective in the downstream classification. In this paper, we propose a Class-Aware Representation rEfinement (CARE) framework for the task of graph classification. CARE computes simple yet powerful class representations and injects them to steer the learning of graph representations towards better class separability. CARE is a plug-and-play framework that is highly flexible and able to incorporate arbitrary GNN backbones without significantly increasing the computational cost. We also theoretically prove that CARE has a better generalization upper bound than its GNN backbone through Vapnik-Chervonenkis (VC) dimension analysis. Our extensive experiments with 11 well-known GNN backbones on 9 benchmark datasets validate the superiority and effectiveness of CARE over its GNN counterparts. Ministry of Education (MOE) National Research Foundation (NRF) Submitted/Accepted version This research/project is supported by the Ministry of Education, Singapore under its MOE Academic Research Fund Tier 2 (STEM RIE2025 Award MOE-T2EP20220-0006), and the National Research Foundation, Singapore under its Industry Alignment Fund – Prepositioning (IAF-PP) Funding Initiative. 2024-10-14T08:48:26Z 2024-10-14T08:48:26Z 2024 Journal Article Xu, J., Ni, J. & Ke, Y. (2024). A class-aware representation refinement framework for graph classification. Information Sciences, 679, 121061-. https://dx.doi.org/10.1016/j.ins.2024.121061 0020-0255 https://hdl.handle.net/10356/180546 10.1016/j.ins.2024.121061 2-s2.0-85196853645 679 121061 en MOE-T2EP20220-0006 IAF-PP Information Sciences © 2024 Elsevier Inc. All rights reserved. This article may be downloaded for personal use only. Any other use requires prior permission of the copyright holder. The Version of Record is available online at http://doi.org/10.1016/j.ins.2024.121061. application/pdf
institution Nanyang Technological University
building NTU Library
continent Asia
country Singapore
Singapore
content_provider NTU Library
collection DR-NTU
language English
topic Computer and Information Science
Graph neural network
Representation learning
Graph classification
spellingShingle Computer and Information Science
Graph neural network
Representation learning
Graph classification
Xu, Jiaxing
Ni, Jinjie
Ke, Yiping
A class-aware representation refinement framework for graph classification
description Graph Neural Networks (GNNs) are widely used for graph representation learning. Despite its prevalence, GNN suffers from two drawbacks in the graph classification task, the neglect of graph-level relationships, and the generalization issue. Each graph is treated separately in GNN message passing/graph pooling, and existing methods to address overfitting operate on each individual graph. This makes the graph representations learnt less effective in the downstream classification. In this paper, we propose a Class-Aware Representation rEfinement (CARE) framework for the task of graph classification. CARE computes simple yet powerful class representations and injects them to steer the learning of graph representations towards better class separability. CARE is a plug-and-play framework that is highly flexible and able to incorporate arbitrary GNN backbones without significantly increasing the computational cost. We also theoretically prove that CARE has a better generalization upper bound than its GNN backbone through Vapnik-Chervonenkis (VC) dimension analysis. Our extensive experiments with 11 well-known GNN backbones on 9 benchmark datasets validate the superiority and effectiveness of CARE over its GNN counterparts.
author2 College of Computing and Data Science
author_facet College of Computing and Data Science
Xu, Jiaxing
Ni, Jinjie
Ke, Yiping
format Article
author Xu, Jiaxing
Ni, Jinjie
Ke, Yiping
author_sort Xu, Jiaxing
title A class-aware representation refinement framework for graph classification
title_short A class-aware representation refinement framework for graph classification
title_full A class-aware representation refinement framework for graph classification
title_fullStr A class-aware representation refinement framework for graph classification
title_full_unstemmed A class-aware representation refinement framework for graph classification
title_sort class-aware representation refinement framework for graph classification
publishDate 2024
url https://hdl.handle.net/10356/180546
_version_ 1814777774616870912