Mutual Information Maximization in Graph Neural Networks

TL;DR

This paper introduces a mutual information maximization approach to enhance graph neural networks, aiming to reduce information loss during aggregation and improve performance across various graph tasks.

Contribution

It proposes enlarging the neighborhood in GNN aggregation to maximize mutual information, which is a novel extension to existing frameworks.

Findings

Improved performance on graph classification tasks

Enhanced results in link prediction and edge classification

Outperforms state-of-the-art methods on benchmark datasets

Abstract

A variety of graph neural networks (GNNs) frameworks for representation learning on graphs have been recently developed. These frameworks rely on aggregation and iteration scheme to learn the representation of nodes. However, information between nodes is inevitably lost in the scheme during learning. In order to reduce the loss, we extend the GNNs frameworks by exploring the aggregation and iteration scheme in the methodology of mutual information. We propose a new approach of enlarging the normal neighborhood in the aggregation of GNNs, which aims at maximizing mutual information. Based on a series of experiments conducted on several benchmark datasets, we show that the proposed approach improves the state-of-the-art performance for four types of graph tasks, including supervised and semi-supervised graph classification, graph link prediction and graph edge generation and classification.