Russian Federation
Russian Federation
The possibilities of artificial neural networks as a basis for creating artificial intelligence systems are considered. The main advantage of artificial neural networks is their ability to learn, generalize, and identify hidden dependencies in the source data. The peculiarities, classification, and methods of training artificial neural networks and the tasks successfully solved using these networks are presented. The stages of solving various problems using artificial neural networks are considered. The features of fuzzy neural networks, deep neural networks, and convolutional neural networks are considered in detail. The following examples of intelligent systems developed on the basis of artificial neural networks are considered: a system for approximating functions from noisy data, a system for classifying large amounts of data, an automatic control system, a cryptographic system, and an image recognition system. Computer models of these systems have been developed, implemented as computer programs. The results of the work of the considered intelligent systems are presented in a visual graphical form.
artificial intelligence, artificial neural networks, fuzzy neural networks, deep neural networks, convolutional neural networks, computer models, function approximation, classification, automatic control, cryptographic system, image recognition
1. Maksimov V.I., Kachaev S.V. Cognitive technologies for management decision support. M.: Proceedings of the IPU, 2018. Vol. 1.
2. Dimitrichenko D.P., Zhidkov R.A. The use of neural networks to solve data processing problems in the construction of cognitive maps // News of the Kabardino-Balkarian scientific center of the russian academy of sciences. 2017. № 6.
3. Anderson C.H. The self-organizing map // Neural Computation. 2012. Vol. 5.
4. Barron A.R. Neural net approximation // IEEE Transaction on Information Theory. 2013. Vol. 49.
5. Vapnik V.N. A fuzzy neural network model // IEEE Transactions on Neural Networks. 2017. Vol. 3. № 5.
6. Labinsky A.Yu. Approximation of functions of many variables by a fuzzy neural network // Natural and man-made risks (physico-mathematical and applied aspects). 2017. № 2. P. 6–14.
7. Labinsky A.Yu. Usage of neural networks for solving classification problems // Natural and man-made risks (physico-mathematical and applied aspects). 2017. № 4. P. 25–32.
8. Labinsky A.Yu. The use of fuzzy logic and neural networks in automatic control systems // Scientific and analytical journal «Vestnik Saint-Petersburg university of State fire service of EMERCOM of Russia». 2019. № 1. P. 11–17.
9. Labinsky A.Yu. Neural networks and information security // Problems of risk management in the technosphere. 2019. № 1. P. 16–24.
10. Labinsky A.Yu. Features of using a neural network for image recognition // Problems of risk management in the technosphere. 2016. № 3. P. 21–27.
