Large-Scale Representation Learning on Graphs via Bootstrapping

TL;DR

This paper introduces BGRL, a scalable self-supervised graph representation learning method that predicts augmented inputs without negative examples, achieving state-of-the-art results on large graphs with reduced memory costs.

Contribution

We propose BGRL, a simple, scalable self-supervised learning approach for graphs that eliminates the need for negative samples and complex augmentations, enabling large-scale applications.

Findings

BGRL outperforms prior methods on multiple benchmarks.

Achieves 2-10x reduction in memory costs.

Successfully scales to graphs with hundreds of millions of nodes.

Abstract

Self-supervised learning provides a promising path towards eliminating the need for costly label information in representation learning on graphs. However, to achieve state-of-the-art performance, methods often need large numbers of negative examples and rely on complex augmentations. This can be prohibitively expensive, especially for large graphs. To address these challenges, we introduce Bootstrapped Graph Latents (BGRL) - a graph representation learning method that learns by predicting alternative augmentations of the input. BGRL uses only simple augmentations and alleviates the need for contrasting with negative examples, and is thus scalable by design. BGRL outperforms or matches prior methods on several established benchmarks, while achieving a 2-10x reduction in memory costs. Furthermore, we show that BGRL can be scaled up to extremely large graphs with hundreds of millions of nodes in the semi-supervised regime - achieving state-of-the-art performance and improving over supervised baselines where representations are shaped only through label information. In particular, our solution centered on BGRL constituted one of the winning entries to the Open Graph Benchmark - Large Scale Challenge at KDD Cup 2021, on a graph orders of magnitudes larger than all previously available benchmarks, thus demonstrating the scalability and effectiveness of our approach.