Equivalence of approximation by convolutional neural networks and fully-connected networks

DOI10.1090/proc/14789OpenAlexW2982376398WikidataQ126860813 ScholiaQ126860813MaRDI QIDQ5218202

Philipp Petersen, Felix Voigtlaender

Publication date: 2 March 2020

Published in: Proceedings of the American Mathematical Society (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/1809.00973

zbMATH Keywords

rate of convergenceneural networksfunction approximationconvolutional neural networks

Mathematics Subject Classification ID

Convolution as an integral transform (44A35) Rate of convergence, degree of approximation (41A25) Approximation by operators (in particular, by integral operators) (41A35) Algorithms for approximation of functions (65D15) Approximation by arbitrary nonlinear expressions; widths and entropy (41A46)

Related Items (19)

Approximation properties of deep ReLU CNNs ⋮ Full error analysis for the training of deep neural networks ⋮ Theory of deep convolutional neural networks. III: Approximating radial functions ⋮ Neural network approximation of continuous functions in high dimensions with applications to inverse problems ⋮ Convergence of deep convolutional neural networks ⋮ On decision regions of narrow deep neural networks ⋮ Universal regular conditional distributions via probabilistic transformers ⋮ A convenient infinite dimensional framework for generative adversarial learning ⋮ Error bounds for approximations using multichannel deep convolutional neural networks with downsampling ⋮ The universal approximation theorem for complex-valued neural networks ⋮ Learning ability of interpolating deep convolutional neural networks ⋮ Approximation bounds for convolutional neural networks in operator learning ⋮ Numerical solution of the parametric diffusion equation by deep neural networks ⋮ Theory of deep convolutional neural networks. II: Spherical analysis ⋮ On the rate of convergence of image classifiers based on convolutional neural networks ⋮ Structure-preserving deep learning ⋮ Optimal Approximation with Sparsely Connected Deep Neural Networks ⋮ Approximating functions with multi-features by deep convolutional neural networks ⋮ A theoretical analysis of deep neural networks and parametric PDEs

Cites Work

• Unnamed Item
• Unnamed Item
• Approximation by superposition of sigmoidal and radial basis functions
• Lower bounds for approximation by MLP neural networks
• Multilayer feedforward networks are universal approximators
• Provable approximation properties for deep neural networks
• Approximation properties of a multilayered feedforward artificial neural network
• Optimal approximation of piecewise smooth functions using deep ReLU neural networks
• Error bounds for approximations with deep ReLU networks
• Universal approximation bounds for superpositions of a sigmoidal function
• Deep distributed convolutional neural networks: Universality
• Optimal Approximation with Sparsely Connected Deep Neural Networks
• Approximation by superpositions of a sigmoidal function

This page was built for publication: Equivalence of approximation by convolutional neural networks and fully-connected networks