Representation theory and invariant neural networks
Representation theory and invariant neural networks
A feedforward neural network is a computational device used for pattern recognition. In many recognition problems, certain transformations exist which, when applied to a pattern, leave its classification unchanged. Invariance under a given group of transformations is therefore typically a desirable property of pattern classifiers. In this paper, we present a methodology, based on representation theory, for the construction of a neural network invariant under any given finite linear group. Such networks show improved generalisation abilities and may also learn faster than corresponding networks without inbuilt invariance. We hope in the future to generalise this theory to approximate invariance under continuous groups.
33-60
Wood, J.
65587872-7126-469a-851a-d60195d39058
Shawe-Taylor, J.
c32d0ee4-b422-491f-8c28-78663851d6db
August 1996
Wood, J.
65587872-7126-469a-851a-d60195d39058
Shawe-Taylor, J.
c32d0ee4-b422-491f-8c28-78663851d6db
Wood, J. and Shawe-Taylor, J.
(1996)
Representation theory and invariant neural networks.
Discrete Applied Mathematics, 69 (1-2), .
(doi:10.1016/0166-218X(95)00075-3).
Abstract
A feedforward neural network is a computational device used for pattern recognition. In many recognition problems, certain transformations exist which, when applied to a pattern, leave its classification unchanged. Invariance under a given group of transformations is therefore typically a desirable property of pattern classifiers. In this paper, we present a methodology, based on representation theory, for the construction of a neural network invariant under any given finite linear group. Such networks show improved generalisation abilities and may also learn faster than corresponding networks without inbuilt invariance. We hope in the future to generalise this theory to approximate invariance under continuous groups.
This record has no associated files available for download.
More information
Published date: August 1996
Organisations:
Electronics & Computer Science
Identifiers
Local EPrints ID: 250474
URI: http://eprints.soton.ac.uk/id/eprint/250474
ISSN: 0166-218X
PURE UUID: 80f3abb8-efe1-4d19-b0ae-703ad4aebb91
Catalogue record
Date deposited: 01 Jun 1999
Last modified: 14 Mar 2024 04:52
Export record
Altmetrics
Contributors
Author:
J. Wood
Author:
J. Shawe-Taylor
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics