Abstract:<p>Learning with graph structured data is of great significance for many practical applications. A crucial and fundamental task in graph learning is node classification. In reality, graph nodes are often encoded with various attributes. In addition, the task is usually multi-labeled in nature. In this paper, we tackle the problem of multi-label graph node classification, by leveraging structure, attribute and label information simultaneously. Specifically, to obtain rational node feature representations, we propose an intuitive yet effective graph convolution module to aggregate local attribute information of a given node. Moreover, the homophily hypothesis motivates us to build a label attention module. By exploiting both input and output contextual representations, we utilize the additive attention mechanism and build a label-aware representation learning framework to measure the compatibility between pairs of node embeddings and label embeddings. The proposed novel neural network-based, multi-label classification method has been verified by extensive experiments conducted on five public-available benchmark datasets, including both attributed and non-attributed networks. The results demonstrate the effectiveness of the proposed model with respect to micro-F1, macro-F1 and Hamming loss, comparing with several state-of-the-art methods, including two relational neighbor classifiers and several popular graph neural network models.</p>

What problem does this paper attempt to address?

The paper primarily proposes a novel solution to the problem of multi-label node classification in graph-structured data. Specifically, the paper addresses the following two main issues: 1. **Effective Representation of Node Features**: In graph-structured data, nodes often have rich attribute information. How to effectively combine these attribute information with the topological structure between nodes to form a compact node representation is a challenge. To solve this problem, the authors propose an intuitive and effective graph convolution module that aggregates local neighborhood information of nodes through one-dimensional convolution operations. 2. **Utilizing Label Information to Improve Multi-Label Learning**: Traditional methods often treat labels as atomic symbols without fully utilizing the information that labels themselves may carry. Based on the homogeneity assumption (i.e., individuals with similar attributes tend to have the same labels), the paper further proposes a label attention module to capture the compatibility between node features and label semantics, thereby better utilizing output context information for node classification. In summary, the main contributions of the paper include: - Proposing a one-dimensional convolution module that can efficiently extract node representation vectors from the local neighborhood of nodes. - Introducing a label attention module that can capture the dependency between node features and label embeddings. - Proposing a neural network-based method (LANC, Label Attentive Neighborhood Convolution) to solve the problem of multi-label node classification on graph-structured data, and validating the effectiveness of this method through experiments. This is the first exploration of the application of label-aware representations in multi-label graph node classification. The experimental part uses 5 publicly available datasets for validation, including both attributed and unattributed network data. The results show that the proposed model outperforms existing advanced methods under various evaluation metrics.

Multi-label graph node classification with label attentive neighborhood convolution

Multi-label Node Classification On Graph-Structured Data

Correlation-Aware Graph Convolutional Networks for Multi-Label Node Classification

Multi-Label Image Recognition With Graph Convolutional Networks

Multi-Label Classification with Label Graph Superimposing

GNN-MultiFix: Addressing the pitfalls for GNNs for multi-label node classification

Graph Attention Transformer Network for Multi-label Image Classification

Graph classification based on structural features of significant nodes and spatial convolutional neural networks

Multi-label remote sensing image classification with deformable convolutions and graph neural networks

Graph Decoupling Attention Markov Networks for Semisupervised Graph Node Classification

Node classification based on Attribute Fuse Edge Features and Label Adaptive Adjustment

Multilabel Classification Based on Graph Neural Networks

Graph-MLP: Node Classification without Message Passing in Graph

Multilabel Recognition Algorithm With Multigraph Structure

A Novel Two-Channel Classification Approach Using Graph Attention Network with K-Nearest Neighbor

Graph Multi-Convolution and Attention Pooling for Graph Classification

Hierarchical Graph Attention Networks for Semi-Supervised Node Classification

Location-aware convolutional neural networks for graph classification

Learning Graph Convolutional Networks for Multi-Label Recognition and Applications

GraphHop: An Enhanced Label Propagation Method for Node Classification

Hierarchical Graph Convolutional Networks for Semi-supervised Node Classification