Result: Sequence processing neural network with a non-monotonic transfer function

Title:
Sequence processing neural network with a non-monotonic transfer function
Authors:
KATAYAMA, Katsuki, HORIGUCHI, Tsuyoshi
Source:
Journal of the Physical Society of Japan. 70(5):1300-1314
Publisher Information:
Tokyo; Tokyo: Institute of Pure and Applied Physics, Physical Society of Japan, 2001.
Publication Year:
2001
Physical Description:
print, 31 ref
Original Material:
INIST-CNRS
Subject Terms:
Crystallography, Cristallographie cristallogenèse, Mechanics acoustics, Mécanique et acoustique, Condensed state physics, Physique de l'état condensé, Theoretical physics, Physique théorique, Sciences exactes et technologie, Exact sciences and technology, Physique, Physics, Generalites, General, Instruments, appareillage, composants et techniques communs à plusieurs branches de la physique et de l'astronomie, Instruments, apparatus, components and techniques common to several branches of physics and astronomy, Informatique en physique expérimentale, Computers in experimental physics, Réseaux neuronaux, logique floue, intelligence artificielle, Neural networks, fuzzy logic, artificial intelligence, Domaines classiques de la physique (y compris les applications), Fundamental areas of phenomenology (including applications), Optique, Optics, Formation des images et traitement optique, Imaging and optical processing, Fonctions de transfert optique et de transfert de modulation, Modulation and optical transfer functions, Etude théorique, Theoretical study, Fonction génératrice, Generating function, Función generatriz, Fonction monotone, Monotonic function, Función monótona, Fonction transfert, Transfer functions, Réseau neuronal, Neural networks, Simulation numérique, Digital simulation, Suite mathématique, Sequence(mathematics), Sucesión matemática
Document Type:
Academic journal Article
File Description:
text
Language:
English
Author Affiliations:
Department of Computer and Mathematical Sciences, Graduate School of Information Sciences, Tohoku University, Sendai 980-8579, Japan
ISSN:
0031-9015
Access URL:
http://pascal-francis.inist.fr/vibad/index.php?action=search&terms=1119957
Rights:
Copyright 2001 INIST-CNRS
CC BY 4.0
Sauf mention contraire ci-dessus, le contenu de cette notice bibliographique peut être utilisé dans le cadre d’une licence CC BY 4.0 Inist-CNRS / Unless otherwise stated above, the content of this bibliographic record may be used under a CC BY 4.0 licence by Inist-CNRS / A menos que se haya señalado antes, el contenido de este registro bibliográfico puede ser utilizado al amparo de una licencia CC BY 4.0 Inist-CNRS
Notes:
Metrology

Physics: optics
Accession Number:
edscal.1119957
Database:
PASCAL Archive

Further Information

We investigate storage capacity and retrieval property for a synchronous fully connected neural network with a non-monotonic transfer function which retrieves sequences of patterns, by an analytic method and also by numerical simulations. Because of asymmetry of interactions and non-monotonicity of the transfer function, it is difficult to use conventional methods of the equilibrium statistical mechanics in order to investigate the network. We then use a generating-function method of path-integral representation, and obtain equations for dynamical order parameters in the stationary state. We clarify that the network with the non-monotonic transfer function retrieves more sequences of patterns than that with a monotonic transfer function at zero temperature when non-monotonicity of the transfer function is selected optimally. It is also clarified that some chaotic behavior appears in solutions for the equations of the dynamical order parameters when non-monotonicity of the transfer function increases. The analytic results are in excellent agreement with the results obtained by numerical simulations.