Neural and Spectral Operator Surrogates on Gaussian Spaces

DOI10.48550/arXiv.2607.08492MaRDI QIDQ7283932arXivFDO

Authors Carlo Marcati, Christoph Schwab, Mario Marić, Jakob Zech

Publication date 9 July 2026

Published in arXiv

Tldr We prove expression rate bounds for spectral and neural surrogates approximating holomorphic maps between separable Hilbert spaces. The novel contribution is an algebraic decay assumption on the eigenvalues of covariance operators which yields convergence rates for mean-square errors and Sobolev norms, both in Gaussian measures and their gradients.

Full work available at URL https://doi.org/10.48550/arXiv.2607.08492

Mathematics Subject Classification ID

Second-order elliptic equations (35J15) Rate of convergence, degree of approximation (41A25) Approximation by arbitrary nonlinear expressions; widths and entropy (41A46) Derivatives of functions in infinite-dimensional spaces (46G05) Infinite-dimensional holomorphy (46G20) Artificial neural networks and deep learning (68T07)

arXiv classification

math.NA

This page was built for publication: Neural and Spectral Operator Surrogates on Gaussian Spaces

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7283932)