From Spectral Methods to Sample Complexity Bounds for Fourier Neural Operators

DOI10.48550/arXiv.2607.00320MaRDI QIDQ7270500arXivFDO

Authors Nisha Chandramoorthy, Daniel Sanz-Alonso, Nathan Waniorek

Publication date 1 July 2026

Published in arXiv

Tldr We establish approximation and learning guarantees for Fourier neural operators applied to time-T solution operators of dissipative evolution equations by connecting spectral methods with sample complexity bounds. We derive polynomial sample complexity guarantees that depend on the smoothness of the input space, dimension of the physical domain, and dissipation strength.

Full work available at URL https://doi.org/10.48550/arXiv.2607.00320

arXiv classification

cs.LGmath.NAstat.ML

This page was built for publication: From Spectral Methods to Sample Complexity Bounds for Fourier Neural Operators

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7270500)