On the approximation of rough functions with deep neural networks

DOI10.1007/S40324-022-00299-WMaRDI QIDQ2089012zbMATH OpenOpenAlexWikidataFDO

Authors Tim De Ryck, Siddhartha Mishra, Deep Ray

Publication date 6 October 2022

Published in S\(\vec{\text{e}}\)MA Journal (Search for Journal in Brave)

Copyright license Creative Commons Attribution 4.0 International

Full work available at URL https://arxiv.org/abs/1912.06732

zbMATH Keywords

data compressiondeep ReLU networksENOinterpolationrough functions

Mathematics Subject Classification ID

Artificial intelligence (68Txx)

Recommendations

• Constructive deep ReLU neural network approximation
• Error bounds for approximations with deep ReLU networks
• Nonlinear approximation and (deep) ReLU networks
• Deep ReLU networks and high-order finite element methods
• Provable approximation properties for deep neural networks

Cites work

• A survey of several finite difference methods for systems of nonlinear hyperbolic conservation laws
• Approximation of piecewise smooth functions and images by edge-adapted (ENO-EA) nonlinear multiresolution techniques
• Deep learning
• Deep learning observables in computational fluid dynamics
• Efficient implementation of essentially nonoscillatory shock-capturing schemes. II
• ENO reconstruction and ENO interpolation are stable
• ENO schemes with subcell resolution
• Error bounds for approximations with deep ReLU networks
• High-Order Essentially Nonoscillatory Schemes for Hamilton–Jacobi Equations
• scientific article; zbMATH DE number 1181255 (Why is no real title available?)
• Hyperbolic Conservation Laws in Continuum Physics
• Interpolation and Approximation of Piecewise Smooth Functions
• Multilayer feedforward networks are universal approximators
• Nonlinear multiscale decompositions: The approach of A. Harten
• Onsager and the theory of hydrodynamic turbulence
• Optimal approximation of piecewise smooth functions using deep ReLU neural networks
• Topological properties of the set of functions generated by neural networks of fixed size
• TVB Runge-Kutta Local Projection Discontinuous Galerkin Finite Element Method for Conservation Laws II: General Framework
• Uniformly high order accurate essentially non-oscillatory schemes. III
• Uniformly high order accurate essentially non-oscillatory schemes. III. (Reprint)
• Universal approximation bounds for superpositions of a sigmoidal function

Cited in

(7)

• Function approximation by deep neural networks with parameters \(\{0, \pm \frac{1}{2}, \pm 1,2\}\)
• Function approximation by deep networks
• On the approximation of rough functions with deep neural networks
• Error analysis for deep neural network approximations of parametric hyperbolic conservation laws
• Runge-Kutta physics informed neural networks: formulation and analysis
• Learning WENO for entropy stable schemes to solve conservation laws
• Deep Ritz--finite element methods: neural network methods trained with finite elements

Describes a project that uses

Uses Software

• Adam

This page was built for publication: On the approximation of rough functions with deep neural networks

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2089012)