Activation functions

N. C. Steele, C. R. Reeves, Elena Gaura

Research output: Chapter in Book/Report/Conference proceedingConference proceeding contribution

80 Citations (Scopus)

Abstract

This paper considers an alternative activation function for use with MLP networks. The performance on parity problems is considered and it has been found that only n - I hidden units were needed to resolve the n-bit problem. Also, insight has been gained into the families of network parameters generated. Use as the kernel of a support vector machine for particular problems is anticipated.

Original languageEnglish
Title of host publicationArtificial Neural Nets and Genetic Algorithms
Subtitle of host publicationProceedings of the International Conference in Prague, Czech Republic, 2001
EditorsVera Kurkova, Nigel C. Steele, Roman Neruda, Miroslav Karny
Place of PublicationNew York
PublisherSpringer, Springer Nature
Pages27-30
Number of pages4
ISBN (Electronic)9783709162309
ISBN (Print)3211836519, 9783211836514
DOIs
Publication statusPublished - 2001
Externally publishedYes
EventInternational Conference on Artificial Neural Networks and Genetic Algorithms (ICANNGA) - Prague, Czech Republic
Duration: 22 Apr 200125 Apr 2001

Publication series

NameSpringer Computer Science
PublisherSpringer-Verlag Wien

Conference

ConferenceInternational Conference on Artificial Neural Networks and Genetic Algorithms (ICANNGA)
CountryCzech Republic
CityPrague
Period22/04/0125/04/01

Cite this

Steele, N. C., Reeves, C. R., & Gaura, E. (2001). Activation functions. In V. Kurkova, N. C. Steele, R. Neruda, & M. Karny (Eds.), Artificial Neural Nets and Genetic Algorithms: Proceedings of the International Conference in Prague, Czech Republic, 2001 (pp. 27-30). (Springer Computer Science). New York: Springer, Springer Nature. https://doi.org/10.1007/978-3-7091-6230-9_5