Mostra el registre d'ítem simple
Neural networks with periodic and monotonic activation functions: a comparative study in classification problems
dc.contributor.author | Romero Merino, Enrique |
dc.contributor.author | Sopena, Josep Maria |
dc.contributor.author | Alquézar Mancho, René |
dc.contributor.author | Moliner, Joan L. |
dc.contributor.other | Universitat Politècnica de Catalunya. Departament de Llenguatges i Sistemes Informàtics |
dc.date.accessioned | 2016-04-01T12:52:01Z |
dc.date.available | 2016-04-01T12:52:01Z |
dc.date.issued | 2000-02 |
dc.identifier.citation | Romero, E., Sopena, J., Alquézar, R., Moliner, J. "Neural networks with periodic and monotonic activation functions: a comparative study in classification problems". 2000. |
dc.identifier.uri | http://hdl.handle.net/2117/85060 |
dc.description.abstract | This article discusses a number of reasons why the use of non-monotonic functions as activation functions can lead to a marked improvement in the performance of a neural network. Using a wide range of benchmarks we show that a multilayer feed-forward network using sine activation functions (and an appropriate choice of initial parameters) learns much faster than one incorporating sigmoid functions - as much as 150-500 times faster - when both types are trained with backpropagation. Learning speed also compares favorably with speeds reported using modified versions of the backpropagation algorithm. In addition, computational and generalization capacity increases. |
dc.format.extent | 16 p. |
dc.language.iso | eng |
dc.relation.ispartofseries | LSI-00-12-R |
dc.subject | Àrees temàtiques de la UPC::Informàtica::Intel·ligència artificial |
dc.subject.other | Non-monotonic functions |
dc.subject.other | Neural networks |
dc.title | Neural networks with periodic and monotonic activation functions: a comparative study in classification problems |
dc.type | External research report |
dc.contributor.group | Universitat Politècnica de Catalunya. SOCO - Soft Computing |
dc.rights.access | Open Access |
local.identifier.drac | 1847205 |
dc.description.version | Postprint (published version) |
local.citation.author | Romero, E.; Sopena, J.; Alquézar, R.; Moliner, J. |
Fitxers d'aquest items
Aquest ítem apareix a les col·leccions següents
-
Reports de recerca [55]
-
Reports de recerca [1.107]