Functionally Equivalent Feedforward Neural Networks

Neural ComputationPublished 1 May 1994

Věra Kůrková, Paul C. Kainen

DOI10.1162/neco.1994.6.3.543

Source URL

Citations60

SJR quartileQ1

SJR score0.83

SNIP1.45

TL;DR

For a feedforward perceptron type architecture with a single hidden layer but with a quite general activation function, the relation between pairs of weight vectors determining networks with the same input-output function is characterized.

Abstract

For a feedforward perceptron type architecture with a single hidden layer but with a quite general activation function, we characterize the relation between pairs of weight vectors determining networks with the same input-output function.

Keywords

Computer Science

Computers in PhysicsTen Lectures on Wavelets

15,386 Citations1992Ingrid Daubechies, Christopher Heil

This paper presents a meta-analyses of the wavelet transforms of Coxeter’s inequality and its applications to multiresolutional analysis and orthonormal bases.

Neural NetworksApproximation capabilities of multilayer feedforward networks

6,086 Citations1991Kurt Hornik

It is shown that standard multilayer feedforward networks with as few as a single hidden layer and arbitrary bounded and nonconstant activation function are universal approximators with respect to L p (μ) performance criteria, for arbitrary finite input environment measures μ.

SIAM ReviewWavelets and Dilation Equations: A Brief Introduction

950 Citations1989Gilbert Strang

It is shown in Part 1 how conditions on the $c_k $ lead to approximation properties and orthogonality properties of the wavelets, and the recursive algorithms that decompose and reconstruct f.

A Survey of Modern Algebra

673 Citations2017Garrett Birkhoff, Saunders Mac Lane

Neural NetworksUniqueness of the weights for minimal feedforward nets with a given input-output map

264 Citations1992Héctor J. Sussmann

It is shown that, for feedforward nets with a single hidden layer, a single output node, and a ''transfer function'' Tanh s, the net is uniquely determined by its input-output map, up to an obvious finite group of symmetries.

Approximating and learning unknown mappings using multilayer feedforward networks with bounded weights

108 Citations1990Maxwell B. Stinchcombe, Hannah J. White

It is shown that feedforward networks having bounded weights are not undesirable restricted, but are in fact universal approximators, provided that the hidden-layer activation function belongs to one of several suitable broad classes of functions: polygonal functions, certain piecewise polynomial functions, or a class of functions analytic on some open interval.

Neural ComputationOn the Geometry of Feedforward Neural Network Error Surfaces

97 Citations1993An Mei Chen, Haw-minn Lu +1 more

The geometric structure of these equioutput weight space transformations is explored for the case of multilayer perceptron networks with tanh activation functions and it is shown that these transformations form an algebraic group isomorphic to a direct product of Weyl groups.

Neural NetworksFor neural networks, function determines form

93 Citations1993Francesca Albertini, Eduardo D. Sontag

Elsevier eBooksON THE ALGEBRAIC STRUCTURE OF FEEDFORWARD NETWORK WEIGHT SPACES

56 Citations1990Robert Hecht-Nielsen

It is shown that the set of invertible linear weight transformations that leave the input/output transfer function of the entire network unchanged form a subgroup of the group of invertedible linear mappings from the weight space to itself.

On the geometry of feedforward neural network weight spaces

13 Citations1991An Mei Chen, Robert Hecht-Nielsen

The authors show that there exists a cone in weight space which forms a minimal sufficient search set for learning and show that the average distance between global minimum copies on a finite sphere known to contain a global minimum goes to zero as the number of weights in the neural network increases without bound.