Semi-Supervised Graph Embedding for Multi-Label Graph Node Classification

TL;DR

This paper introduces ML-GCN, a semi-supervised graph neural network that embeds nodes and labels in a shared space, capturing label correlations to improve multi-label node classification performance.

Contribution

The paper proposes a novel GCN-based approach that embeds labels and nodes in a unified space and models label correlations, enhancing multi-label classification accuracy.

Findings

ML-GCN outperforms four state-of-the-art methods on several datasets.

The approach effectively captures label-label and node-label correlations.

Experimental results demonstrate improved classification performance.

Abstract

The graph convolution network (GCN) is a widely-used facility to realize graph-based semi-supervised learning, which usually integrates node features and graph topologic information to build learning models. However, as for multi-label learning tasks, the supervision part of GCN simply minimizes the cross-entropy loss between the last layer outputs and the ground-truth label distribution, which tends to lose some useful information such as label correlations, so that prevents from obtaining high performance. In this paper, we pro-pose a novel GCN-based semi-supervised learning approach for multi-label classification, namely ML-GCN. ML-GCN first uses a GCN to embed the node features and graph topologic information. Then, it randomly generates a label matrix, where each row (i.e., label vector) represents a kind of labels. The dimension of the label vector is the same as that of the node vector before the last convolution operation of GCN. That is, all labels and nodes are embedded in a uniform vector space. Finally, during the ML-GCN model training, label vectors and node vectors are concatenated to serve as the inputs of the relaxed skip-gram model to detect the node-label correlation as well as the label-label correlation. Experimental results on several graph classification datasets show that the proposed ML-GCN outperforms four state-of-the-art methods.