Towards Generalizable Graph Contrastive Learning: An Information Theory Perspective

TL;DR

This paper introduces a theoretical framework and new metrics for assessing and improving the generalization ability of graph contrastive learning, leading to state-of-the-art results on benchmarks.

Contribution

It proposes a mutual information-based generalization metric for GCL, provides a theoretical upper bound, and develops the InfoAdv framework to enhance GCL generalization.

Findings

GCL-GE effectively measures GCL generalization ability.

InfoAdv achieves state-of-the-art performance on benchmarks.

Theoretical bounds guide the design of better GCL models.

Abstract

Graph contrastive learning (GCL) emerges as the most representative approach for graph representation learning, which leverages the principle of maximizing mutual information (InfoMax) to learn node representations applied in downstream tasks. To explore better generalization from GCL to downstream tasks, previous methods heuristically define data augmentation or pretext tasks. However, the generalization ability of GCL and its theoretical principle are still less reported. In this paper, we first propose a metric named GCL-GE for GCL generalization ability. Considering the intractability of the metric due to the agnostic downstream task, we theoretically prove a mutual information upper bound for it from an information-theoretic perspective. Guided by the bound, we design a GCL framework named InfoAdv with enhanced generalization ability, which jointly optimizes the generalization metric and InfoMax to strike the right balance between pretext task fitting and the generalization ability on downstream tasks. We empirically validate our theoretical findings on a number of representative benchmarks, and experimental results demonstrate that our model achieves state-of-the-art performance.