SCL-GNN: Towards Generalizable Graph Neural Networks via Spurious Correlation Learning

TL;DR

This paper introduces SCL-GNN, a novel graph neural network framework that improves generalization by identifying and mitigating spurious correlations using HSIC and bi-level optimization, outperforming existing methods under distribution shifts.

Contribution

The paper proposes SCL-GNN, a new GNN framework that explicitly learns and reduces spurious correlations to enhance robustness and generalization across IID and OOD graph data.

Findings

SCL-GNN outperforms state-of-the-art methods on real-world datasets.

The framework effectively mitigates spurious correlations.

Experiments show improved robustness under distribution shifts.

Abstract

Graph Neural Networks (GNNs) have demonstrated remarkable success across diverse tasks. However, their generalization capability is often hindered by spurious correlations between node features and labels in the graph. Our analysis reveals that GNNs tend to exploit imperceptible statistical correlations in training data, even when such correlations are unreliable for prediction. To address this challenge, we propose the Spurious Correlation Learning Graph Neural Network (SCL-GNN), a novel framework designed to enhance generalization on both Independent and Identically Distributed (IID) and Out-of-Distribution (OOD) graphs. SCL-GNN incorporates a principled spurious correlation learning mechanism, leveraging the Hilbert-Schmidt Independence Criterion (HSIC) to quantify correlations between node representations and class scores. This enables the model to identify and mitigate irrelevant but influential spurious correlations effectively. Additionally, we introduce an efficient bi-level optimization strategy to jointly optimize modules and GNN parameters, preventing overfitting. Extensive experiments on real-world and synthetic datasets demonstrate that SCL-GNN consistently outperforms state-of-the-art baselines under various distribution shifts, highlighting its robustness and generalization capabilities.