Mostrar el registro sencillo del ítem
Study of the Effect of Combining Activation Functions in a Convolutional Neural Network
dc.contributor.author | Vergara Villegas, Osslan Osiris | |
dc.date.accessioned | 2020-12-10T16:49:15Z | |
dc.date.available | 2020-12-10T16:49:15Z | |
dc.date.issued | 2020-11-04 | es_MX |
dc.identifier.uri | http://cathi.uacj.mx/20.500.11961/15659 | |
dc.description.abstract | Convolutional Neural Networks (CNN’s) have proven to be an effective approach for solving image classification problems. The output, the accuracy and the computational efficiency of a CNN are determined mainly by the architecture, the convolutional filters, and the activation functions. Based on the importance of an activation function, in this paper, nine new activation functions based on combinations of classical functions such as ReLU and sigmoid are presented. Also, a study about the effects caused by the activation functions in the performance of a CNN is presented. First, every new function is described, also, their graphs, analytic forms and derivatives are presented. Then, a traditional CNN model with each new activation function is used to classify three 10-class databases: MNIST, Fashion MNIST and a handwritten digit database created by us. Experimental results illustrate that some of the proposed activation functions lead to better performances on classifying than classical activation functions. Moreover, our study demonstrated that the accuracy of a CNN could be increased by 1.18% with the new proposed activation functions. | es_MX |
dc.description.uri | https://latamt.ieeer9.org/index.php/transactions/article/view/4134 | es_MX |
dc.language.iso | en | es_MX |
dc.relation.ispartof | Producto de investigación IIT | es_MX |
dc.relation.ispartof | Instituto de Ingeniería y Tecnología | es_MX |
dc.subject | Activation function | es_MX |
dc.subject | Convolutional neural network | es_MX |
dc.subject | Modified National Institute of Standards and Technology | es_MX |
dc.subject | Fashion Modified National Institute of Standards and Technology | es_MX |
dc.subject | Sigmoid | es_MX |
dc.subject | Rectified linear unit | es_MX |
dc.subject.other | info:eu-repo/classification/cti/7 | es_MX |
dc.title | Study of the Effect of Combining Activation Functions in a Convolutional Neural Network | es_MX |
dc.type | Artículo | es_MX |
dcterms.thumbnail | http://ri.uacj.mx/vufind/thumbnails/rupiiit.png | es_MX |
dcrupi.instituto | Instituto de Ingeniería y Tecnología | es_MX |
dcrupi.cosechable | Si | es_MX |
dcrupi.norevista | 1e | es_MX |
dcrupi.volumen | 100 | es_MX |
dcrupi.nopagina | 1-9 | es_MX |
dc.contributor.coauthor | Cruz Sanchez, Vianey Guadalupe | |
dc.contributor.coauthor | Ochoa Domínguez, Humberto | |
dc.contributor.coauthor | Nandayapa, Manuel | |
dc.contributor.alumno | 171517 | es_MX |
dc.journal.title | IEEE Latin America Transactions | es_MX |
dc.lgac | VISIÓN, INSTRUMENTACIÓN Y CONTROL | es_MX |
dc.cuerpoacademico | Visión Artificial, Control y Robótica | es_MX |
dc.contributor.authorexterno | Sossa Azuela, Juan Humberto |