INCREASING INNATE ROBUSTNESS IN ARTIFICIAL NEURAL NETWORKS USING REDUNDANCY

Authors
THOMPSON MP KAMBHAMPATI C
Citation
Mp. Thompson et C. Kambhampati, INCREASING INNATE ROBUSTNESS IN ARTIFICIAL NEURAL NETWORKS USING REDUNDANCY, Electronics Letters, 31(22), 1995, pp. 1931-1932
Citations number
5
Categorie Soggetti
Engineering, Eletrical & Electronic
Journal title
Electronics Letters → ACNP
ISSN journal
00135194
Volume
31
Issue
22
Year of publication
1995
Pages
1931 - 1932
Database
ISI
SICI code
0013-5194(1995)31:22<1931:IIRIAN>2.0.ZU;2-F
Abstract
A theoretical explanation of robustness and its relationship with redu nduncy is proposed and used to derive a novel and powerful technique w hich allows the innate robustness of most types of artificial neural n etwork (ANN) to be enhanced to a user-defined degree.