Locality preserving dense graph convolutional networks with graph context-aware node representations

DOI10.1016/J.NEUNET.2021.05.031MaRDI QIDQ6055073zbMATH OpenOpenAlexFDO

Authors Wenfeng Liu, Maoguo Gong, Zedong Tang, A. K. Qin, Kai Sheng, Mingliang Xu

Publication date 28 September 2023

Published in Neural Networks (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/2010.05404

zbMATH Keywords

dense connectiongraph classificationgraph convolutional networklocality preservationnode representationrepresentation learning

Mathematics Subject Classification ID

Graph theory (including graph drawing) in computer science (68R10) Learning and adaptive systems in artificial intelligence (68T05) Pattern recognition, speech recognition (68T10)

Abstract: Graph convolutional networks (GCNs) have been widely used for representation learning on graph data, which can capture structural patterns on a graph via specifically designed convolution and readout operations. In many graph classification applications, GCN-based approaches have outperformed traditional methods. However, most of the existing GCNs are inefficient to preserve local information of graphs -- a limitation that is especially problematic for graph classification. In this work, we propose a locality-preserving dense GCN with graph context-aware node representations. Specifically, our proposed model incorporates a local node feature reconstruction module to preserve initial node features into node representations, which is realized via a simple but effective encoder-decoder mechanism. To capture local structural patterns in neighbourhoods representing different ranges of locality, dense connectivity is introduced to connect each convolutional layer and its corresponding readout with all previous convolutional layers. To enhance node representativeness, the output of each convolutional layer is concatenated with the output of the previous layer's readout to form a global context-aware node representation. In addition, a self-attention module is introduced to aggregate layer-wise representations to form the final representation. Experiments on benchmark datasets demonstrate the superiority of the proposed model over state-of-the-art methods in terms of classification accuracy.

Recommendations

• Pseudoinverse graph convolutional networks. Fast filters tailored for large eigengaps of dense graphs and hypergraphs
• Fast Haar transforms for graph neural networks
• Two-order approximate spectral convolutional model for semi-supervised classification
• Attentional multilabel learning over graphs: a message passing approach
• Graph clustering with graph capsule network

Cites work

• \(k\)-hop graph neural networks
• A gentle introduction to deep learning for graphs
• Propagation kernels: efficient graph kernels from propagated information
• Weisfeiler-Lehman graph kernels

Cited in

(13)

• Fast Haar transforms for graph neural networks
• Interpretable multi-scale graph descriptors via structural compression
• Attentional multilabel learning over graphs: a message passing approach
• Two-order approximate spectral convolutional model for semi-supervised classification
• Probabilistic learning on graphs via contextual architectures
• Multi-graph fusion graph convolutional networks with pseudo-label supervision
• Robust node embedding against graph structural perturbations
• Local2global: a distributed approach for scaling representation learning on graphs
• Relation-aware Graph Contrastive Learning
• Joint hyperbolic and Euclidean geometry contrastive graph neural networks
• Informative pseudo-labeling for graph neural networks with few labels
• Saliency-aware regularized graph neural network
• Geometric localized graph convolutional network for multi-view semi-supervised classification

This page was built for publication: Locality preserving dense graph convolutional networks with graph context-aware node representations

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6055073)