ADAPTATION OF THE BACK PROPAGATION ERROR ALGORITHM FOR A CONVOLUTIONAL NEURAL NETWORK WITH SECOND-ORDER NEURONS AND DYNAMIC RECEPTIVE FIELDS
Abstract
The article suggests the adaptation of the back propagation error algorithm for a convolutional neural network with dynamic receptive fields and second-order neurons. A pattern recognition experiments shows that the second-order neurons combination and dynamic receptive fields allows to reduce the generalization error are described.
About the Authors
Roman Mikhailovich Nemkov
North-Caucasus Federal University
Russian Federation
Russian Federation
Oksana Stanislavovna Mezentseva
North-Caucasus Federal University
Russian Federation
Russian Federation
Dmitriy Viktorovich Mezentsev
North-Caucasus Federal University
Russian Federation
Russian Federation
References
1. Назаров Л. Е. Нейросетевые алгоритмы обнаружения, классификации и распознавания объектов на изображениях / Л. Е. Назаров, Н. С. Томашевич, А. Н. Балухто // Нейрокомпьютеры в прикладных задачах обработки изображений. Кн. 25. 2007. С. 25-54.
2. Nemkov R., Mezentseva O., Mezentsev D. Using of a Convolutional Neural Network with Changing Receptive Fields in the Tasks of Image Recognition / Proceedings of the First International Scientific Conference “Intelligent Information Technologies for Industry” (IITI’16), Volume 451 of the series Advances in Intelligent Systems and Computing. pp. 15-23.
3. Лагунов Н. А. Выделение и распознавание объектов с использованием оптимизированного алгоритма селективного поиска и сверточной нейронной сети высокого порядка // Фундаментальные исследования. 2015. №5. С. 511-516.
4. Немков Р М. Метод синтеза параметров математической модели сверточной нейронной сети с расширенным обучающим множеством // Современные проблемы науки и образования, 2015. № 1. URL: http://www.science-education.ru/125-19867/
5. NYU Object Recognition Benchmark (NORB) [электронный ресурс] // URL: www.cs.nyu.edu/~yldab/data/norb-vL0/ (дата обращения 12.12.2016).
6. Nemkov R. Dynamical Change of the Perceiving Properties of Neural Networks as Training with Noise and Its Impact on Pattern Recognition // Young Scientists’ International Workshop on Trends in Information Processing (YSIP). 2014. URL: http://ceur-ws.org/ Vol-1145/paper4.pdf
7. Bengio Y.: Learning deep architectures for AI. Foundations and Trends in Machine Learning, vol. 2, issue 1, 2009,. pp. 1-127.
8. Nemkov R. M. Synthesis method of mathematical model parameters of the convolutional neural network with extended training set. URL: http://www.science-education.ru/125-19867/ (30.01.2016).
9. Nemkov R., Mezentseva O. The Use of Convolutional Neural Networks with Non-specific Receptive Fields. The 4th International Scientific Conference: Applied Natural Sciences. Novy Smokovec, 2013. pp. 284-289.
10. Nemkov R. M., Mezentseva O. S. Dynamical change of the perceiving properties of convolutional neural networks and its impact on generalization. Neurocomputers: development and application, 2015, no. 2, pp. 12-18.
Review
For citations:
Nemkov R.M., Mezentseva O.S., Mezentsev D.V. ADAPTATION OF THE BACK PROPAGATION ERROR ALGORITHM FOR A CONVOLUTIONAL NEURAL NETWORK WITH SECOND-ORDER NEURONS AND DYNAMIC RECEPTIVE FIELDS. Modern Science and Innovations. 2018;(2):50-55. (In Russ.)
Views:
109
Email this article 