当前位置: X-MOL 学术arXiv.math.NA › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Spectral Barron space and deep neural network approximation
arXiv - MATH - Numerical Analysis Pub Date : 2023-09-02 , DOI: arxiv-2309.00788
Yulei Liao, Pingbing Ming

We prove the sharp embedding between the spectral Barron space and the Besov space. Given the spectral Barron space as the target function space, we prove a dimension-free result that if the neural network contains $L$ hidden layers with $N$ units per layer, then the upper and lower bounds of the $L^2$-approximation error are $\mathcal{O}(N^{-sL})$ with $0 < sL\le 1/2$, where $s$ is the smoothness index of the spectral Barron space.

中文翻译:

谱巴伦空间和深度神经网络近似

我们证明了谱巴伦空间和贝索夫空间之间的尖锐嵌入。给定谱巴伦空间作为目标函数空间,我们证明了一个无维结果,即如果神经网络包含 $L$ 隐藏层,每层有 $N$ 个单元,则 $L^2$ 的上限和下限- 近似误差为 $\mathcal{O}(N^{-sL})$,其中 $0 < sL\le 1/2$,其中 $s$ 是谱巴伦空间的平滑度指数。
更新日期:2023-09-06
down
wechat
bug