Nonlinear approximation and (deep) ReLU networks

DOI10.1007/s00365-021-09548-zzbMath1501.41003arXiv1905.02199MaRDI QIDQ2117331

Ingrid Daubechies, Guergana Petrova, Simon Foucart, Ronald A. DeVore, B. Hanin

Publication date: 21 March 2022

Published in: Constructive Approximation (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/1905.02199

zbMATH Keywords

neural networks; expressiveness; approximation power; rectified linear unit (ReLU)

Mathematics Subject Classification ID

68T07: Artificial neural networks and deep learning

92B20: Neural networks for/in biological studies, artificial life and related topics

82C32: Neural nets applied to problems in time-dependent statistical mechanics

41A25: Rate of convergence, degree of approximation

41A30: Approximation by other special function classes

41A46: Approximation by arbitrary nonlinear expressions; widths and entropy

Uses Software

• ImageNet
• GNMT
• AlexNet