Exploring the impact of adaptive rewiring in Graph Neural Networks

2026-02-11

AI Summary

This work addresses the challenges of high memory and computational costs, as well as stringent reliability requirements, faced by graph neural networks (GNNs) when applied to large-scale graphs such as those in N-1 security assessment of power grids. To this end, the authors propose a dynamic sparsification approach that integrates concepts from network science and machine learning. The method combines Erdős–Rényi-based sparsification, an adaptive rewiring mechanism, and an early-stopping strategy to dynamically refine the graph structure during training. Experimental results demonstrate that moderate sparsification enhances model generalization, while adaptive rewiring substantially improves reliability and scalability in safety-critical tasks, effectively balancing representational capacity with computational efficiency.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Natural Language Processing: Safety and Robustness
• Reasoning under Uncertainty: Graphical Models

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Security and Privacy: Large-scale security measurements
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines

Abstract

This paper explores sparsification methods as a form of regularization in Graph Neural Networks (GNNs) to address high memory usage and computational costs in large-scale graph applications. Using techniques from Network Science and Machine Learning, including Erd\H{o}s-R\'enyi for model sparsification, we enhance the efficiency of GNNs for real-world applications. We demonstrate our approach on N-1 contingency assessment in electrical grids, a critical task for ensuring grid reliability. We apply our methods to three datasets of varying sizes, exploring Graph Convolutional Networks (GCN) and Graph Isomorphism Networks (GIN) with different degrees of sparsification and rewiring. Comparison across sparsification levels shows the potential of combining insights from both research fields to improve GNN performance and scalability. Our experiments highlight the importance of tuning sparsity parameters: while sparsity can improve generalization, excessive sparsity may hinder learning of complex patterns. Our adaptive rewiring approach, particularly when combined with early stopping, proves promising by allowing the model to adapt its connectivity structure during training. This research contributes to understanding how sparsity can be effectively leveraged in GNNs for critical applications like power grid reliability analysis.

Problem

• Graph Neural Networks
• sparsification
• computational efficiency
• memory usage
• scalability

Innovation

• adaptive rewiring
• graph sparsification
• graph neural networks
• regularization
• power grid reliability