Implementación de redes neuronales utilizando dispositivos lógicos programables
Fecha
Autor corporativo
Título de la revista
ISSN de la revista
Título del volumen
Editor
Universidad Distrital Francisco José de Caldas
Compartir
Director
Altmetric
Resumen
Descripción
Desde el comienzo de la revolución computacional hubo interés por las máquinas inteligentes. Después del fracaso de los métodos de búsqueda de inteligencia artificial aplicados a problemas reales, un nuevo enfoque: inteligencia computacional retomó el camino basándose en estructuras copiadas de la naturaleza. Las redes neuronales (artificiales) emulan de manera simplificada el funcionamiento de las redes biológicas. La investigación ha estado dirigida primordialmente a simulación en computador de los algoritmos que se proponen, lo que se aleja del comportamiento real de las neuronas biológicas. Este artículo muestra un desarrollo en el que se implementó un sistema para diseñar, simular, crear e interactuar con redes neuronales en un dispositivo físico, para acercarse al concepto biológico. El dispositivo de trabajo (un FPGA) es un arreglo de elemento lógicos, cuyos puntos de interconexión pueden ser programados, lo que modela mejor la evolución de un cerebr biológico.
From the beginning of the computational revolution there was interest for the intelligent machines. After the failure of the searching methods or Artificial Intelligence applied to real problems, a new approach, Computational Intelligence, recaptured the road basing on structures copied from nature. The (artificial) neural networks emulate in a simplified way the operation of the biological networks. Research has been directed primarily to computer simulation of the proposed algorithms, which moves away from the real behavior of the biological neurons. This article shows a development of a system for designing, simulating, creating, and interacting with neural netwoks implemented in a physical device, to come closer to the biological concept. The target device, a FPGA, is an arrangement of logical elements, whose interconnection points can be programmed, wich better models the evolutionof a biological brain.
From the beginning of the computational revolution there was interest for the intelligent machines. After the failure of the searching methods or Artificial Intelligence applied to real problems, a new approach, Computational Intelligence, recaptured the road basing on structures copied from nature. The (artificial) neural networks emulate in a simplified way the operation of the biological networks. Research has been directed primarily to computer simulation of the proposed algorithms, which moves away from the real behavior of the biological neurons. This article shows a development of a system for designing, simulating, creating, and interacting with neural netwoks implemented in a physical device, to come closer to the biological concept. The target device, a FPGA, is an arrangement of logical elements, whose interconnection points can be programmed, wich better models the evolutionof a biological brain.
Palabras clave
Computational intelligence, neuronal networks design, neural networks implementation, bio-inspired systems, FPGA, VHDL, Inteligencia computacional, diseño de redes neuronales, implementación de redes neuronales, sistemas bio-inspirados