Neural network approximation

R DeVore, B Hanin, G Petrova - Acta Numerica, 2021 - cambridge.org
Neural networks (NNs) are the method of choice for building learning algorithms. They are
now being investigated for other numerical tasks such as solving high-dimensional partial …

Deep ReLU neural networks in high-dimensional approximation

D Dũng - Neural Networks, 2021 - Elsevier
We study the computation complexity of deep ReLU (Rectified Linear Unit) neural networks
for the approximation of functions from the Hölder–Zygmund space of mixed smoothness …

Simultaneous approximation of a smooth function and its derivatives by deep neural networks with piecewise-polynomial activations

D Belomestny, A Naumov, N Puchkin, S Samsonov - Neural Networks, 2023 - Elsevier
This paper investigates the approximation properties of deep neural networks with
piecewise-polynomial activation functions. We derive the required depth, width, and sparsity …

Function and derivative approximation by shallow neural networks

Y Li, S Lu - arXiv preprint arXiv:2407.05078, 2024 - arxiv.org
We investigate a Tikhonov regularization scheme specifically tailored for shallow neural
networks within the context of solving a classic inverse problem: approximating an unknown …

Computation complexity of deep ReLU neural networks in high-dimensional approximation

D Dũng, VK Nguyen, MX Thao - arXiv preprint arXiv:2103.00815, 2021 - arxiv.org
The purpose of the present paper is to study the computation complexity of deep ReLU
neural networks to approximate functions in H\" older-Nikol'skii spaces of mixed smoothness …

Collocation approximation by deep neural ReLU networks for parametric elliptic PDEs with lognormal inputs

D Dũng - arXiv preprint arXiv:2111.05504, 2021 - arxiv.org
We obtained convergence rates of the collocation approximation by deep ReLU neural
networks of solutions to elliptic PDEs with lognormal inputs, parametrized by $\boldsymbol …

Approximation theory of tree tensor networks: Tensorized univariate functions

M Ali, A Nouy - Constructive Approximation, 2023 - Springer
We study the approximation of univariate functions by combining tensorization of functions
with tensor trains (TTs)—a commonly used type of tensor networks (TNs). Lebesgue L p …

Differentiable neural networks with repu activation: with applications to score estimation and isotonic regression

G Shen, Y Jiao, Y Lin, J Huang - arXiv preprint arXiv:2305.00608, 2023 - arxiv.org
We study the properties of differentiable neural networks activated by rectified power unit
(RePU) functions. We show that the partial derivatives of RePU neural networks can be …

Theoretical Insights into CycleGAN: Analyzing Approximation and Estimation Errors in Unpaired Data Generation

L Sun, D Shen, H Feng - arXiv preprint arXiv:2407.11678, 2024 - arxiv.org
In this paper, we focus on analyzing the excess risk of the unpaired data generation model,
called CycleGAN. Unlike classical GANs, CycleGAN not only transforms data between two …

Deep ReLU neural network approximation in Bochner spaces and applications to parametric PDEs

D Dũng, DT Pham - Journal of Complexity, 2023 - Elsevier
We investigate non-adaptive methods of deep ReLU neural network approximation in
Bochner spaces L 2 (U∞, X, μ) of functions on U∞ taking values in a separable Hilbert …