Degree of approximation results for feedforward networks approximating unknown mappings and their derivatives

Authors:
Kurt Hornik;Maxwell Stinchcombe;Halbert White;Peter Auer
Affiliations:
Technische Universität, Vienna, Austria;University of California, San Diego, CA USA;University of California, San Diego, CA USA;Technische Universität, Graz, Austria
Venue:
Neural Computation
Year:
1994

Citing 4
Cited 5

Real and complex analysis, 3rd ed.

Real and complex analysis, 3rd ed.
Universal approximation of an unknown mapping and its derivatives using multilayer feedforward networks

Neural Networks
Approximation capabilities of multilayer feedforward networks

Neural Networks
Convergence rates for single hidden layer feedforward networks

Neural Networks

Extraction of Logical Rules from Data by Means of Piecewise-Linear Neural Networks

DS '02 Proceedings of the 5th International Conference on Discovery Science
Air pollutant emissions prediction by process modelling - Application in the iron and steel industry in the case of a re-heating furnace

Environmental Modelling & Software
A review of wave-net identical learning & filling-in in a decomposition space of (JPG-JPEG) sampled images

Artificial Intelligence Review
Self-organizing multilayer perceptron

IEEE Transactions on Neural Networks
Set Membership identification of nonlinear systems

Automatica (Journal of IFAC)

Quantified Score

Hi-index	0.00

Visualization

Abstract

Recently Barron (1993) has given rates for hidden layer feedforward networks with sigmoid activation functions approximating a class of functions satisfying a certain smoothness condition. These rates do not depend on the dimension of the input space. We extend Barron's results to feedforward networks with possibly nonsigmoid activation functions approximating mappings and their derivatives simultaneously. Our conditions are similar but not identical to Barron's, but we obtain the same rates of approximation, showing that the approximation error decreases at rates as fast as n-1/2, where n is the number of hidden units. The dimension of the input space appears only in the constants of our bounds.