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

2026-03-09

AI Summary

Graph Neural Networks (GNNs) are susceptible to spurious correlations between node features and labels in training data, which impairs their generalization performance both in-distribution (IID) and out-of-distribution (OOD). To address this issue, this work proposes SCL-GNN, a novel framework that, for the first time, leverages the Hilbert-Schmidt Independence Criterion (HSIC) to quantify spurious associations between graph representations and class scores. SCL-GNN employs an end-to-end trainable bilevel optimization mechanism to jointly learn debiased representations and GNN parameters. Extensive experiments on multiple real-world and synthetic graph datasets demonstrate that SCL-GNN consistently outperforms existing methods across various distribution shift settings, exhibiting superior generalization capability and robustness.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Reasoning under Uncertainty: Graphical Models
• Search and Optimization: Learning to Search

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies

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.

Problem

• Graph Neural Networks
• Spurious Correlation
• Generalization
• Out-of-Distribution
• Node Features

Innovation

• Spurious Correlation Learning
• Graph Neural Networks
• Hilbert-Schmidt Independence Criterion
• Out-of-Distribution Generalization
• Bi-level Optimization