Self-Supervised Graph Learning with Proximity-based Views and Channel Contrast

TL;DR

This paper introduces a self-supervised graph learning method that enhances node representations by creating additional views based on feature and structure similarity, using contrastive learning to improve expressiveness and efficiency.

Contribution

It proposes a novel view generation technique for graphs and a channel-level contrastive approach, reducing computational costs while capturing long-distance node similarities.

Findings

Effective on diverse graph types

Improves representation quality

Reduces computational complexity

Abstract

We consider graph representation learning in a self-supervised manner. Graph neural networks (GNNs) use neighborhood aggregation as a core component that results in feature smoothing among nodes in proximity. While successful in various prediction tasks, such a paradigm falls short of capturing nodes' similarities over a long distance, which proves to be important for high-quality learning. To tackle this problem, we strengthen the graph with two additional graph views, in which nodes are directly linked to those with the most similar features or local structures. Not restricted by connectivity in the original graph, the generated views allow the model to enhance its expressive power with new and complementary perspectives from which to look at the relationship between nodes. Following a contrastive learning approach, we propose a method that aims to maximize the agreement between representations across generated views and the original graph. We also propose a channel-level contrast approach that greatly reduces computation cost, compared to the commonly used node level contrast, which requires computation cost quadratic in the number of nodes. Extensive experiments on seven assortative graphs and four disassortative graphs demonstrate the effectiveness of our approach.