Dimensional Reweighting Graph Convolutional Networks

TL;DR

This paper introduces DrGCN, a flexible dimensional reweighting method for GCNs that improves stability and performance on various node classification tasks, including large-scale industrial datasets.

Contribution

The paper proposes DrGCN, a novel dimensional reweighting approach that enhances GCN stability and performance, compatible with existing sampling and aggregation techniques.

Findings

DrGCN improves stability of GCNs as proven by mean field theory.

Experimental results show superior performance on benchmark datasets.

Outperforms existing models on Alibaba industrial dataset.

Abstract

Graph Convolution Networks (GCNs) are becoming more and more popular for learning node representations on graphs. Though there exist various developments on sampling and aggregation to accelerate the training process and improve the performances, limited works focus on dealing with the dimensional information imbalance of node representations. To bridge the gap, we propose a method named Dimensional reweighting Graph Convolution Network (DrGCN). We theoretically prove that our DrGCN can guarantee to improve the stability of GCNs via mean field theory. Our dimensional reweighting method is very flexible and can be easily combined with most sampling and aggregation techniques for GCNs. Experimental results demonstrate its superior performances on several challenging transductive and inductive node classification benchmark datasets. Our DrGCN also outperforms existing models on an industrial-sized Alibaba recommendation dataset.