Improving the Long-Range Performance of Gated Graph Neural Networks

TL;DR

This paper introduces a new GNN architecture designed to better capture long-range dependencies in multi-relational graphs, addressing vanishing gradient issues common in existing models.

Contribution

A novel Gated Graph Neural Network architecture that enhances long-range dependency modeling in multi-relational graphs.

Findings

Outperforms existing GNN models on synthetic tasks

Demonstrates improved handling of long-range dependencies

Addresses vanishing gradient problems in GNNs

Abstract

Many popular variants of graph neural networks (GNNs) that are capable of handling multi-relational graphs may suffer from vanishing gradients. In this work, we propose a novel GNN architecture based on the Gated Graph Neural Network with an improved ability to handle long-range dependencies in multi-relational graphs. An experimental analysis on different synthetic tasks demonstrates that the proposed architecture outperforms several popular GNN models.