Constructive deep ReLU neural network approximation

DOI10.1007/S10915-021-01718-2zbMath1500.41002OpenAlexW4205370523MaRDI QIDQ2067309

Joost A. A. Opschoor, Lukas Herrmann, Christoph Schwab

Publication date: 18 January 2022

Published in: Journal of Scientific Computing (Search for Journal in Brave)

Full work available at URL: https://doi.org/10.1007/s10915-021-01718-2

zbMATH Keywords

exponential convergencegeneralization errordeep ReLU neural networksneural network construction

Mathematics Subject Classification ID

Artificial neural networks and deep learning (68T07) Best approximation, Chebyshev systems (41A50) Numerical interpolation (65D05) Approximation by polynomials (41A10) Algorithms for approximation of functions (65D15)

Related Items (7)

Three ways to solve partial differential equations with neural networks — A review ⋮ Optimal approximation of infinite-dimensional holomorphic functions ⋮ Neural and spectral operator surrogates: unified construction and expression rate bounds ⋮ Neural network expression rates and applications of the deep parametric PDE method in counterparty credit risk ⋮ Deep ReLU networks and high-order finite element methods. II: Chebyšev emulation ⋮ Parametric shape holomorphy of boundary integral operators with applications ⋮ Approximation results for gradient flow trained shallow neural networks in \(1d\)

Uses Software

• DeepONet

Cites Work

• Unnamed Item
• Deep learning observables in computational fluid dynamics
• Chebyshev interpolation for parametric option pricing
• Approximation properties of a multilayered feedforward artificial neural network
• Exponential convergence of the deep neural network approximation for analytic functions
• Non-convergence of stochastic gradient descent in the training of deep neural networks
• Shape holomorphy of the Calderón projector for the Laplacian in \(\mathbb{R}^2\)
• B-PINNs: Bayesian physics-informed neural networks for forward and inverse PDE problems with noisy data
• Optimal approximation of piecewise smooth functions using deep ReLU neural networks
• Higher order quasi-Monte Carlo integration for Bayesian PDE inversion
• Error bounds for approximations with deep ReLU networks
• Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
• Auswertung der Normen von Interpolationsoperatoren
• Multilevel higher-order quasi-Monte Carlo Bayesian estimation
• Electromagnetic wave scattering by random surfaces: Shape holomorphy
• Shape Holomorphy of the Stationary Navier--Stokes Equations
• Deep learning in high dimension: Neural network expression rates for generalized polynomial chaos expansions in UQ
• The Gap between Theory and Practice in Function Approximation with Deep Neural Networks
• Optimal Approximation with Sparsely Connected Deep Neural Networks
• Multilevel Quasi-Monte Carlo Uncertainty Quantification for Advection-Diffusion-Reaction
• Convergence rates of high dimensional Smolyak quadrature
• Deep ReLU networks and high-order finite element methods
• Deep neural network expression of posterior expectations in Bayesian PDE inversion
• Better Approximations of High Dimensional Smooth Functions by Deep Neural Networks with Rectified Power Units
• Solving inverse problems using data-driven models
• Well-Posed Bayesian Inverse Problems: Priors with Exponential Tails
• Deep ReLU neural networks in high-dimensional approximation

This page was built for publication: Constructive deep ReLU neural network approximation