Year: 2022
Author: Yan Meng, Pingbing Ming
Communications in Computational Physics, Vol. 32 (2022), Iss. 5 : pp. 1361–1400
Abstract
We introduce a new function space, dubbed as the Barron spectrum space, which arises from the target function space for the neural network approximation. We give a Bernstein type sufficient condition for functions in this space, and clarify the embedding among the Barron spectrum space, the Bessel potential space, the Besov space and the Sobolev space. Moreover, the unexpected smoothness and the decaying behavior of the radial functions in the Barron spectrum space have been investigated. As an application, we prove a dimension explicit $L^q$ error bound for the two-layer neural network with the Barron spectrum space as the target function space, the rate is dimension independent.
You do not have full access to this article.
Already a Subscriber? Sign in as an individual or via your institution
Journal Article Details
Publisher Name: Global Science Press
Language: English
DOI: https://doi.org/10.4208/cicp.OA-2022-0151
Communications in Computational Physics, Vol. 32 (2022), Iss. 5 : pp. 1361–1400
Published online: 2022-01
AMS Subject Headings: Global Science Press
Copyright: COPYRIGHT: © Global Science Press
Pages: 40
Keywords: Fourier transform Besov space Sobolev space radial function neural network.
Author Details
-
Local randomized neural networks with hybridized discontinuous Petrov–Galerkin methods for Stokes–Darcy flows
Dang, Haoning | Wang, FeiPhysics of Fluids, Vol. 36 (2024), Iss. 8
https://doi.org/10.1063/5.0218131 [Citations: 0] -
A finite difference method for elliptic equations with the variable-order fractional derivative
Shi, Siyuan | Hao, Zhaopeng | Du, RuiNumerical Algorithms, Vol. (2024), Iss.
https://doi.org/10.1007/s11075-024-01922-9 [Citations: 0]