Integral Representations of Sobolev Spaces via ReLU$^k$ Activation Function and Optimal Error Estimates for Linearized Networks

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Liu, Xinliang, Mao, Tong, Xu, Jinchao
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916731337834496
author Liu, Xinliang
Mao, Tong
Xu, Jinchao
author_facet Liu, Xinliang
Mao, Tong
Xu, Jinchao
contents This paper presents two main theoretical results concerning shallow neural networks with ReLU$^k$ activation functions. We establish a novel integral representation for Sobolev spaces, showing that every function in $H^{\frac{d+2k+1}{2}}(Ω)$ can be expressed as an $L^2$-weighted integral of ReLU$^k$ ridge functions over the unit sphere. This result mirrors the known representation of Barron spaces and highlights a fundamental connection between Sobolev regularity and neural network representations. Moreover, we prove that linearized shallow networks -- constructed by fixed inner parameters and optimizing only the linear coefficients -- achieve optimal approximation rates $O(n^{-\frac{1}{2}-\frac{2k+1}{2d}})$ in Sobolev spaces.
format Preprint
id arxiv_https___arxiv_org_abs_2505_00351
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Integral Representations of Sobolev Spaces via ReLU$^k$ Activation Function and Optimal Error Estimates for Linearized Networks
Liu, Xinliang
Mao, Tong
Xu, Jinchao
Numerical Analysis
This paper presents two main theoretical results concerning shallow neural networks with ReLU$^k$ activation functions. We establish a novel integral representation for Sobolev spaces, showing that every function in $H^{\frac{d+2k+1}{2}}(Ω)$ can be expressed as an $L^2$-weighted integral of ReLU$^k$ ridge functions over the unit sphere. This result mirrors the known representation of Barron spaces and highlights a fundamental connection between Sobolev regularity and neural network representations. Moreover, we prove that linearized shallow networks -- constructed by fixed inner parameters and optimizing only the linear coefficients -- achieve optimal approximation rates $O(n^{-\frac{1}{2}-\frac{2k+1}{2d}})$ in Sobolev spaces.
title Integral Representations of Sobolev Spaces via ReLU$^k$ Activation Function and Optimal Error Estimates for Linearized Networks
topic Numerical Analysis
url https://arxiv.org/abs/2505.00351