University of Hertfordshire

Paying Attention to Relevant Dimensions: A Localist Approach

Research output: Chapter in Book/Report/Conference proceedingChapter (peer-reviewed)

Documents

View graph of relations
Original languageEnglish
Title of host publicationConnectionist models of learning, development and evolution
Subtitle of host publicationProceedings of the 6th Neural Computation and Psychology Workshop
EditorsRobert M French, Jacques P Sougne
PublisherSpringer
Pages105-112
Number of pages8
ISBN (Print)1852333545, 9781852333546
Publication statusPublished - 2001
Event6th Neural Computation and Psychology Workshop - Liege, Belgium
Duration: 16 Sep 200018 Sep 2000

Workshop

Workshop6th Neural Computation and Psychology Workshop
CountryBelgium
CityLiege
Period16/09/0018/09/00

Abstract

Localist models of, for example, the classification of multidimensional stimuli, can run into problems if generalization is attempted when many of the stimulus dimensions are irrelevant to the classification task in hand. A procedure is suggested by representations that focus on the relevant dimensions only. These permit good generalization which would be lacking in a simple exemplar-based model.

Notes

Michael Page, ‘Paying Attention to Relevant Dimensions: A Localist Approach’, in Robert M. French and Jacques P. Sougne, eds., Connectionist models of learning, development and evolution, Proceedings of the 6th Neural Computation and Psychology Workshop, 16 – 18 September 2000, Liege, Belgium. ISBN 1852333545.

ID: 531282