Handling dropout probability estimation in convolution neural networks using meta-heuristics
dc.contributor.author | Rosa, Gustavo H. de [UNESP] | |
dc.contributor.author | Papa, Joao P. [UNESP] | |
dc.contributor.author | Yang, Xin-S | |
dc.contributor.institution | Universidade Estadual Paulista (Unesp) | |
dc.contributor.institution | Middlesex Univ | |
dc.date.accessioned | 2018-11-26T17:55:04Z | |
dc.date.available | 2018-11-26T17:55:04Z | |
dc.date.issued | 2018-09-01 | |
dc.description.abstract | Deep learning-based approaches have been paramount in recent years, mainly due to their outstanding results in several application domains, ranging from face and object recognition to handwritten digit identification. Convolutional neural networks (CNNs) have attracted a considerable attention since they model the intrinsic and complex brain working mechanisms. However, one main shortcoming of such models concerns their overfitting problem, which prevents the network from predicting unseen data effectively. In this paper, we address this problem by means of properly selecting a regularization parameter known as dropout in the context of CNNs using meta-heuristic-driven techniques. As far as we know, this is the first attempt to tackle this issue using this methodology. Additionally, we also take into account a default dropout parameter and a dropout-less CNN for comparison purposes. The results revealed that optimizing dropout-based CNNs is worthwhile, mainly due to the easiness in finding suitable dropout probability values, without needing to set new parameters empirically. | en |
dc.description.affiliation | Sao Paulo State Univ, Dept Comp, BR-17033360 Bauru, SP, Brazil | |
dc.description.affiliation | Middlesex Univ, Sch Sci & Technol, London NW4 4BT, England | |
dc.description.affiliationUnesp | Sao Paulo State Univ, Dept Comp, BR-17033360 Bauru, SP, Brazil | |
dc.description.sponsorship | Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) | |
dc.description.sponsorship | Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) | |
dc.description.sponsorshipId | FAPESP: 2015/25739-4 | |
dc.description.sponsorshipId | FAPESP: 2014/12236-1 | |
dc.description.sponsorshipId | FAPESP: 2014/16250-9 | |
dc.description.sponsorshipId | CNPq: 306166/2014-3 | |
dc.format.extent | 6147-6156 | |
dc.identifier | http://dx.doi.org/10.1007/s00500-017-2678-4 | |
dc.identifier.citation | Soft Computing. New York: Springer, v. 22, n. 18, p. 6147-6156, 2018. | |
dc.identifier.doi | 10.1007/s00500-017-2678-4 | |
dc.identifier.file | WOS000442576400018.pdf | |
dc.identifier.issn | 1432-7643 | |
dc.identifier.uri | http://hdl.handle.net/11449/164566 | |
dc.identifier.wos | WOS:000442576400018 | |
dc.language.iso | eng | |
dc.publisher | Springer | |
dc.relation.ispartof | Soft Computing | |
dc.relation.ispartofsjr | 0,593 | |
dc.rights.accessRights | Acesso aberto | |
dc.source | Web of Science | |
dc.subject | Convolutional neural networks | |
dc.subject | Dropout | |
dc.subject | Meta-heuristic optimization | |
dc.title | Handling dropout probability estimation in convolution neural networks using meta-heuristics | en |
dc.type | Artigo | |
dcterms.license | http://www.springer.com/open+access/authors+rights?SGWID=0-176704-12-683201-0 | |
dcterms.rightsHolder | Springer | |
unesp.campus | Universidade Estadual Paulista (Unesp), Faculdade de Ciências, Bauru | pt |
unesp.department | Computação - FC | pt |
Arquivos
Pacote Original
1 - 1 de 1
Carregando...
- Nome:
- WOS000442576400018.pdf
- Tamanho:
- 1.32 MB
- Formato:
- Adobe Portable Document Format
- Descrição: