Title :
Implementation of neural constructivism with programmable hardware
Author :
Perez-Uribe, A. ; Sanchez, E.
Author_Institution :
Logic Syst. Lab., Swiss Federal Inst. of Technol., Lausanne
Abstract :
Most neural network models base their “learning” capability on changing the strengths of interconnection between computational elements. However, according to “neural constructivism”, an environmentally-guided neural circuit building offers powerful learning capabilities while minimizing the need for domain-specific structure prespecification. This paper presents a field programmable hardware implementation of an unsupervised constructive neural network with online size adaptation, a form of neural constructivism, and presents a color learning and recognition application
Keywords :
feedforward neural nets; image colour analysis; image segmentation; learning systems; neural net architecture; unsupervised learning; color recognition; constructive learning; feedforward neural nets; field programmable hardware; image segmentation; neural constructivism; unsupervised constructive neural network; Biological neural networks; Buildings; Concurrent computing; Field programmable gate arrays; Integrated circuit interconnections; Network topology; Neural network hardware; Neural networks; Neurons; Power system modeling;
Conference_Titel :
Neuro-Fuzzy Systems, 1996. AT'96., International Symposium on
Conference_Location :
Lausanne
Print_ISBN :
0-7803-3367-5
DOI :
10.1109/ISNFS.1996.603820
