Functional and Structural Features of Recurrent Neural Networks with Controlled Elements
The features of recurrent neural networks with controlled elements are considered. The functions of these networks for controlled associative processing of signals are refined. A number of the space-time structures of such networks are analyzed. Among them, the neural networks with one-, two-, and three-level structures of layers are investigated. The results of studies on the stability of the associative processing of distorted signals by these networks are reflected. Based on the simulation results, recommendations are formulated to expand the possibilities of associative signal processing in recurrent neural networks with controlled elements.
KeywordsRecurrent neural network Logical structure Associative signal processing Control
This research is partially supported by the RFBR foundation grant No 16-29-09482.
- 1.Haykin, S.: Neural Networks and Learning Machines, 3rd edn. Prentice Hall, New-York (2008)Google Scholar
- 3.Hopfield, J.J.: Neural networks and physical systems with emergent collective computational abilities. Proc. Nat. Acad. Sci. USA 79(8), 2554–2558 (1982)Google Scholar
- 4.Kosko, B.: Bidirectional associative memories. IEEE Trans. Syst. Man Cybern. 18(1), 49–60 (1988)Google Scholar
- 5.Hawkins, J., Blakeslee, S.: On Intelligence. Brown Walker (2006)Google Scholar
- 6.Palm, G.: Neural associative memories and sparse coding. Neural Networks 37, 165–171 (2013)Google Scholar
- 7.Osipov, V.Y.: Associative intellectual machine. Inf. Technol. Comput. Mach. 2, 59–67 (2010)Google Scholar
- 8.Osipov, V., Osipova, M.: Method and device of intellectual processing of information in neural network. RU Patent No 2413304 (2011)Google Scholar
- 10.Osipov, V.: Structure and basic functions of cognitive neural network machine. In: 12th International Scientific-Technical Conference on Electromechanics and Robotics, Zavalishin’s Readings, pp. 1–5 (2017)Google Scholar
- 12.Osipov, V., Osipova, M.: Space-time signal binding in recurrent neural networks with controlled elements. Neurocomputing 308, 194–204 (2018)Google Scholar