Browsing by Autor "Beatriz A. Garro"

Now showing 1 - 9 of 9

An Evolutionary Feature-Based Visual Attention Model Applied to Face Recognition
(Springer Science+Business Media, 2010) Roberto A. Vázquez; Humberto Sossa; Beatriz A. Garro
Automatic Design of Artificial Neural Networks by means of Differential Evolution (DE) Algorithm
(2012) Beatriz A. Garro; Humberto Sossa; Roberto A. Vázquez
Classification of DNA Microarrays Using Artificial Bee Colony (ABC) Algorithm
(Springer Science+Business Media, 2014) Beatriz A. Garro; Roberto A. Vázquez; Katya Rodríguez
Design of Artificial Neural Networks Using Differential Evolution Algorithm
(Springer Science+Business Media, 2010) Beatriz A. Garro; Humberto Sossa; Roberto A. Vázquez
Diseño automático de redes neuronales artificiales mediante el uso del algoritmo de evolución diferencial (ED)
(2012) Beatriz A. Garro; Humberto Sossa; Roberto A. Vázquez
Resumen-En el rea de la Inteligencia Artificial, las Redes Neuronales Artificiales (RNA) han sido aplicadas para la solucin de m ltiples tareas. A pesar de su declive y del resurgimiento de su desarrollo y aplicacin, su dise o se ha caracterizado por un mecanismo de prueba y error, el cual puede originar un desempe o bajo. Por otro lado, los algoritmos de aprendizaje que se utilizan como el algoritmo de retropropagacin y otros basados en el gradiente descenciente, presentan una desventaja: no pueden resolver problemas no continuos ni problemas multimodales. Por esta razn surge la idea de aplicar algoritmos evolutivos para disear de manera automtica una RNA. En esta investigacin, el algoritmo de Evolucin Diferencial (ED) encuentra los mejores elementos principales de una RNA: la arquitectura, los pesos sinpticos y las funciones de transferencia. Por otro lado, dos funciones de aptitud son propuestas: el error cuadrtico medio (MSE por sus siglas en ingls) y el error de clasificacin (CER) las cuales, involucran la etapa de validacin para garantizar un buen desempe o de la RNA. Primero se realiz un estudio de las diferentes configuraciones del algoritmo de ED, y al determinar cul fue la mejor configuracin se realiz una experimentacin exhaustiva para medir el desempeo de la metodologa propuesta al resolver problemas de
EEG Channel Selection using Fractal Dimension and Artificial Bee Colony Algorithm
(2018) Beatriz A. Garro; R. Salazar-Varas; Roberto A. Vázquez
The development of Brain Computer Interfaces (BCI) has attracted the attention of several research groups for solving different kind of problems in the field of medicine, device control, gaming, etc. However, there are several challenges that must be solved in order to have more feasible BCI applications. One of these challenges is related with the high dimensionality of the EEG recordings due to they are acquired from several channels to preserve high spatial accuracy. However, it is necessary to carefully select the channels that provide the most relevant information as well as the feature extraction technique in order to guarantee an acceptable accuracy. On the other hand, Swarm Intelligence techniques are based on biological processes of some species for their survival such as the food search task or natural selection process. These techniques have been widely applied in several optimizations problems obtaining acceptable results. In this paper, we described the high dimensionality challenge as an optimization problem in order to applied a swarm optimization technique, the Artificial Bee Colony (ABC) algorithm, to determine the set of channels that are more useful to discriminate different mental tasks. Furthermore, during the feature extraction stage, we applied fractal dimension methods to build the feature vector that will be used to train a classifier. Finally, the accuracy of the proposed methodology is tested classifying different motor tasks using the data set IVa from BCI international competition III.
Evolving Neural Networks: A Comparison between Differential Evolution and Particle Swarm Optimization
(Springer Science+Business Media, 2011) Beatriz A. Garro; Humberto Sossa; Roberto A. Vázquez
Generalized neurons and its application in DNA microarray classification
(2016) Beatriz A. Garro; Katya Rodríguez; Roberto A. Vázquez
The DNA Microarray classification is an important task in bioinformatics and medicine area. The genetic expression in DNA microarrays present the opportunity to determine for example, which genes are involved with a particular disease, identify tumors, select the best treatment, etc. Several computational intelligence technique such as artificial neural networks can be used to identify different groups of genes associated with a particular disease. However, the enormous quantity of genes and the few samples available demand the use of more robust artificial neural networks. The purpose of this research is focused on showing how a generalize neuron (GN) can be applied in the DNA microarray classification task. In order to do that, the proposed methodology first, select the set of genes that best describe the disease applying the artificial bee colony algorithm. After that, the genes found during the first stage are used to train a GN. The GN is trained with the differential evolution algorithm. Finally, the accuracy of the proposed methodology is tested classifying two type of cancer using DNA microarrays: the acute lymphocytic leukemia and the acute myeloid leukemia.
Training Spiking Neurons by Means of Particle Swarm Optimization
(Springer Science+Business Media, 2011) Roberto A. Vázquez; Beatriz A. Garro