N-GCN: Multi-scale Graph Convolution for Semi-supervised Node Classification

TL;DR

N-GCN introduces a multi-scale graph convolution model that combines supervised GCNs with unsupervised random walk embeddings, significantly improving semi-supervised node classification performance across multiple datasets.

Contribution

The paper proposes N-GCN, a novel multi-scale GCN framework that integrates random walk-based embeddings with supervised learning, enhancing classification accuracy and robustness.

Findings

Outperforms state-of-the-art on Cora, Citeseer, Pubmed, and PPI datasets.

Generalizes to semi-supervised methods like GraphSAGE, leading to N-SAGE.

Shows resilience to adversarial input perturbations.

Abstract

Graph Convolutional Networks (GCNs) have shown significant improvements in semi-supervised learning on graph-structured data. Concurrently, unsupervised learning of graph embeddings has benefited from the information contained in random walks. In this paper, we propose a model: Network of GCNs (N-GCN), which marries these two lines of work. At its core, N-GCN trains multiple instances of GCNs over node pairs discovered at different distances in random walks, and learns a combination of the instance outputs which optimizes the classification objective. Our experiments show that our proposed N-GCN model improves state-of-the-art baselines on all of the challenging node classification tasks we consider: Cora, Citeseer, Pubmed, and PPI. In addition, our proposed method has other desirable properties, including generalization to recently proposed semi-supervised learning methods such as GraphSAGE, allowing us to propose N-SAGE, and resilience to adversarial input perturbations.