Radiotekhnika
Publishing house Radiotekhnika

"Publishing house Radiotekhnika":
scientific and technical literature.
Books and journals of publishing houses: IPRZHR, RS-PRESS, SCIENCE-PRESS


Тел.: +7 (495) 625-9241

 

Hardware implementation of a neuron-like network with the ability to recognize noisy pattern

Keywords:

I.A. Mishustin – Post-graduate Student, Institute of Precision Mechanics and Computer Engineering n.a. S.A. Lebedev RAS (Moscow)
E-mail: mishustinivan777@gmail.com
N.B. Preobrazhensky – Ph.D. (Eng.), Senior Research Scientist, Institute of Precision Mechanics and Com-puter Engineering n.a. S.A. Lebedev RAS (Moscow)
E-mail: nbp@mail.ru
A.A. Zhdanov – Dr.Sc. (Phis.-Math.), Professor, Institute of Precision Mechanics and Computer Engineering n.a. S.A. Lebedev RAS (Moscow)
E-mail: a.zhdanov@mail.ru
I.V. Stepanyan – Dr.Sc. (Biol.), Ph.D. (Eng.), Laboratory of biomechanic systems of Institute of Machine
Science named after A.A. Blagonravov of the Russian Academy of Sciences,
Research Institute of Occupational Health named after N.F. Izmerov of the Russian Academy of Sciences,
Moscow State Сonservatory named after P.I. Chaikovsky
E-mail: neurocomp.pro@gmail.com


An approach to constructing a neuron-like network based on a biologically inspired method of "autonomous adaptive control" (AAC) is considered. The results of the investigation of the methods for constructing the model of a neuron and a neuron-like network described in the AAC method with consideration of a number of specific properties are presented. The main result of this work is the implementation in the language of the description of Verilog equipment presented in the method of AAC of a biologically inspired model of a self-taught neuron possessing noise immunity, the possibility of forgetting and retraining. With the help of developed Neurox software tool specially designed for this project, examples of networks from such neurons are collected, which makes it possible to move on to the hardware neuron-like implementation of the AAC on the FPGA as a whole. Development became possible as a result of close scientific cooperation of co-authors with the support of CJSC "Intellect" (Moscow).

References:
  1. Zhdanov A.A. Avtonomnyj iskusstvennyj intellekt. Izd. 2-e. M.: BINOM. Laboratoriya znanij. 2009. 359 s. (Adaptivnye i intellektual'nye sistemy).
  2. Kryzhanovskij M.V. Principy nejronopodobnoj realizacii sistem Avtonomnogo Adaptivnogo Upravleniya: Dis. ... kand. fiz.-mat. nauk. M. 2004. 129 c.
  3. Zhdanov A.A. Application of Pattern Recognition Procedure to the Acquisition and Use of Data in Control // Pattern Recognition and Image Analysis. 1992. V. 2. № 2. Р. 180–194.
  4. Zhdanov A.A. A principle of Pattern Formation and Recognition // Pattern Recognition and Image Analysis. 1992. V. 2. № 3. P. 249–264.
  5. Zhdanov A.A., Ryadovikov A.V. Neuron Models in the Autonomous Adaptive Control Method // Optical Memory and Neural Network, Allerton Press, Inc. 2000. V. 9. № 2. P. 115–132.
  6. Zhdanov A.A. Metod avtonomnogo adaptivnogo upravleniya // Izv. Akademii Nauk. Teoriya i sistemy upravleniya. 1999. № 5. S. 127–134.
  7. Zhdanov A.A., Preobrazhenskij N.B., Holopov YU.A. Stepanyan I.V., Nguen Hyu CHung Apparatnaya realizaciya nejronnoj seti v adaptivnoj sisteme upravleniya // Nejrokomp'yutery: razrabotka, primenenie. 2016. № 6. S. 55–62.
  8. Dileep George, Jeff Hawkins. Towards a Mathematical Theory of Cortical Micro-circuits. October 9, 2009 PLoS Computational Biology, Edited by Karl J. Friston, vol. 5, issue 10, p. e1000532
  9. http://neurox.intellect-labs.com/
  10. Karshenbojm I. Kratkij kurs HDL CHast' 5. Napisanie koda nezavisimogo ot apparatnoj platformy // Komponenty i tekhnologii 2008. № 9.

© Издательство «РАДИОТЕХНИКА», 2004-2017            Тел.: (495) 625-9241                   Designed by [SWAP]Studio