Residual or Gate? Towards Deeper Graph Neural Networks for Inductive Graph Representation Learning

TL;DR

This paper introduces a recurrent graph neural network (RGNN) that uses recurrent units to enable deeper GNNs, effectively capturing long-term dependencies and reducing noise, leading to state-of-the-art results on multiple benchmarks.

Contribution

The paper proposes a novel RGNN architecture with recurrent units to facilitate deeper GNNs and improve node representation learning.

Findings

Achieved state-of-the-art results on Pubmed, Reddit, and PPI datasets.

Recurrent units help prevent noisy information in graphs.

Deeper GNNs are possible with the proposed method.

Abstract

In this paper, we study the problem of node representation learning with graph neural networks. We present a graph neural network class named recurrent graph neural network (RGNN), that address the shortcomings of prior methods. By using recurrent units to capture the long-term dependency across layers, our methods can successfully identify important information during recursive neighborhood expansion. In our experiments, we show that our model class achieves state-of-the-art results on three benchmarks: the Pubmed, Reddit, and PPI network datasets. Our in-depth analyses also demonstrate that incorporating recurrent units is a simple yet effective method to prevent noisy information in graphs, which enables a deeper graph neural network.