Estimating condition number with Graph Neural Networks

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Carson, Erin, Chen, Xinye
Format: Preprint
Publié: 2026
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866917344265109504
author Carson, Erin
Chen, Xinye
author_facet Carson, Erin
Chen, Xinye
contents For large sparse matrices, we almost never compute the condition number exactly because that would require computing the full SVD or full eigenvalue decompositionIn this paper, we propose a fast method for estimating the condition number of sparse matrices using graph neural networks (GNNs). To enable efficient training and inference of GNNs, our proposed feature engineering for GNNs achieves $\mathrm{O}(\mathrm{nnz} + n)$, where $\mathrm{nnz}$ is the number of non-zero elements in the matrix and $n$ denotes the matrix dimension. We propose two prediction schemes for estimating the matrix condition number using GNNs. One follows by decomposing the condition number and predicts the relatively more computationally intensive part $\|\mathbf{A}^{-1}\|$, while the other is to predict the whole condition number $κ$. Our approach can be extended to an arbitrary norm. The extensive experiments for the two schemes are conducted for 1-norm and 2-norm condition number estimation, which show that our method achieves a significant speedup over the traditional numerical estimation methods.
format Preprint
id arxiv_https___arxiv_org_abs_2603_10277
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Estimating condition number with Graph Neural Networks
Carson, Erin
Chen, Xinye
Machine Learning
Numerical Analysis
For large sparse matrices, we almost never compute the condition number exactly because that would require computing the full SVD or full eigenvalue decompositionIn this paper, we propose a fast method for estimating the condition number of sparse matrices using graph neural networks (GNNs). To enable efficient training and inference of GNNs, our proposed feature engineering for GNNs achieves $\mathrm{O}(\mathrm{nnz} + n)$, where $\mathrm{nnz}$ is the number of non-zero elements in the matrix and $n$ denotes the matrix dimension. We propose two prediction schemes for estimating the matrix condition number using GNNs. One follows by decomposing the condition number and predicts the relatively more computationally intensive part $\|\mathbf{A}^{-1}\|$, while the other is to predict the whole condition number $κ$. Our approach can be extended to an arbitrary norm. The extensive experiments for the two schemes are conducted for 1-norm and 2-norm condition number estimation, which show that our method achieves a significant speedup over the traditional numerical estimation methods.
title Estimating condition number with Graph Neural Networks
topic Machine Learning
Numerical Analysis
url https://arxiv.org/abs/2603.10277