Logo do repositório
 

κ-Entropy Based Restricted Boltzmann Machines

dc.contributor.authorPassos, Leandro Aparecido
dc.contributor.authorCleison Santana, Marcos [UNESP]
dc.contributor.authorMoreira, Thierry [UNESP]
dc.contributor.authorPapa, Joao Paulo [UNESP]
dc.contributor.institutionUniversidade Federal de São Carlos (UFSCar)
dc.contributor.institutionUniversidade Estadual Paulista (Unesp)
dc.date.accessioned2020-12-12T02:27:01Z
dc.date.available2020-12-12T02:27:01Z
dc.date.issued2019-07-01
dc.description.abstractRestricted Boltzmann Machines achieved notorious popularity in the scientific community in the last decade due to outstanding results in a wide range of applications and also for providing the required mechanisms to build successful deep learning models, i.e., Deep Belief Networks and Deep Boltzmann Machines. However, their main bottleneck is related to the learning step, which is usually time-consuming. In this paper, we introduce a Sigmoid-like family of functions based on the Kaniadakis entropy formulation in the context of the RBM learning procedure. Experiments concerning binary image reconstruction are conducted in four public datasets to evaluate the robustness of the proposed approach. The results suggest that such a family of functions is suitable to increase the convergence rate when compared to standard functions employed by the research community.en
dc.description.affiliationDepartment of Computing UFSCar - Federal University of São Carlos
dc.description.affiliationSchool of Sciences UNESP - São Paulo State University
dc.description.affiliationUnespSchool of Sciences UNESP - São Paulo State University
dc.identifierhttp://dx.doi.org/10.1109/IJCNN.2019.8851714
dc.identifier.citationProceedings of the International Joint Conference on Neural Networks, v. 2019-July.
dc.identifier.doi10.1109/IJCNN.2019.8851714
dc.identifier.scopus2-s2.0-85073157165
dc.identifier.urihttp://hdl.handle.net/11449/201217
dc.language.isoeng
dc.relation.ispartofProceedings of the International Joint Conference on Neural Networks
dc.sourceScopus
dc.subjectKaniadakis Entropy
dc.subjectMachine Learning
dc.subjectRestricted Boltzmann Machines
dc.titleκ-Entropy Based Restricted Boltzmann Machinesen
dc.typeTrabalho apresentado em eventopt
dspace.entity.typePublication
relation.isDepartmentOfPublication872c0bbb-bf84-404e-9ca7-f87a0fe94e58
relation.isDepartmentOfPublication.latestForDiscovery872c0bbb-bf84-404e-9ca7-f87a0fe94e58
relation.isOrgUnitOfPublicationaef1f5df-a00f-45f4-b366-6926b097829b
relation.isOrgUnitOfPublication.latestForDiscoveryaef1f5df-a00f-45f4-b366-6926b097829b
unesp.campusUniversidade Estadual Paulista (UNESP), Faculdade de Ciências, Baurupt
unesp.departmentComputação - FCpt

Arquivos