4 resultados
Resultados de la búsqueda
Mostrando 1 - 4 de 4
Ítem Ambiente software de entrenamiento de redes neuronales con ajuste evolutivo de la topología y las funciones de activación(Universidad Icesi, 2009-08-18T21:03:41Z) Méndez Ortiz, Edgar; Mariño, Juan Sebastián; Arguello Fuentes, HenryThis research examines two problems in the optimization in the neural networks used for most real applications: first, architectural design that involves determining the number of layers and neurons by layer, and second, the activation functions that will be should use in each of these layers. For it is developed a software tool based on genetic algorithms to find these parameters of a neural network. The developed tool allows the user to choose the algorithm used for training and also apply techniques to achieve better generalization such as the early stopping, the repetition of training and adjusting the training data to the activation functions used. Finally, the developed tool is tested into a specialized group of users who use the tool to find an optimal neural network architecture to solve a problem of identity verification through the facial image using artificial neural networks.Ítem Public-Resource Computing: Un nuevo paradigma para la computación y la ciencia(Universidad Icesi, 2006-10-11) Osorio Betancur, Juan DavidThis article explores the Public-Resource Computing concept, an idea that has been successfully developed in the scientific community a few years ago, which involves the use of spare computational resources available in the millions of PCs in the world connected to Internet. We discuss the SETI@home project, the most successful example of this concept, and we describe the BOINC (Berkeley Open Infrastructure for Network Computing) platform, a software system created to make it easy for scientist to create and operate public-resource computing projects.Ítem Combining spectral and fractal features for emotion recognition on Electroencephalographic signals(World Scientific and Engineering Academy and Society, 2014-01-01) Ulloa Villegas, Gonzalo VicenteRecent studies have attempted to recognize emotions by extracting spectral and fractal features from electroencephalographic signals; however, up to now none of them have combined these two features to recognize emotions. This paper aims at providing a comparison between an accuracy rate of an approach that recognizes emotions by extracting both spectral and fractal features with that of those that extract only one of these features. To this end, we designed and implemented a procedure that recognizes positive and negative emotions by extracting spectral, fractal, or both features. Next, using this procedure, we built three different approaches to recognize positive and negative emotions; the first one extracted both spectral and fractal features, whereas the other two extracted each type of feature separately.Ítem Fundamentos y perspectivas de algoritmos inspirados en sistemas biológicos(Universidad Icesi, 2015-01-01) Barona Figueroa, Mónica; Quintero Moreno, Fernando; Asesor TesisDesde la década de 1940 se hizo una primera apuesta de sugerir la computación natural para el desarrollo de algoritmos en la resolución de diversos problemas complejos. Desde entonces, a la medida que se desarrolló la programación matemática y las cualidades computacionales hicieron posible las soluciones de modelos que dan soporte a decisiones organizacionales, se fue teniendo como referente a los sistemas naturales, contenidos dentro la biomímesis, para desarrollar algoritmos, inspirados biológicamente, que brinden soluciones a problemas más complejos, a través de la creación de métodos meta heurísticos, análogos a funcionalidades de redes neuronales, procesos evolutivos, sistemas sociales enjambres y al sistema inmunológico, todo dado a que la programación matemática queda imposibilitada en dar solución en tiempos razonables para problemas combinatorios. Es entonces que el objetivo de este trabajo parte del interés de conocer meta heurísticas inspiradas en la biología, partiendo del origen de la biomímesis y conociendo sus múltiples posibilidades metafóricas que sirven de fuente de inspiración en la búsqueda de soluciones reales en ingeniería.