Zhurnal Radioelektroniki - Journal of Radio Electronics. eISSN 1684-1719. 2023. №2
ContentsFull text in Russian (pdf)
DOI: https://doi.org/10.30898/1684-1719.2023.2.7
USING GENERATIVE-ADVERSARIAL NEURAL NETWORKS
TO EXPAND TRAINING (TEST) SELECTIONS
M.L. Belokopytov
Military space academy by A. Mozhaysky name
197198, Russia, St. Petersburg, Zhdanovskaya st., 13
The paper was received December 21, 2022.
Abstract. The article is devoted to the formation of the initial data set for training neural networks. The main methods of formation and expansion of training (test) samples are presented, in particular with the help of augmentation, mathematical modeling methods, as well as generative-adversarial neural networks. The principle of operation of GAN networks is considered, the algorithm of operation of the generator and discriminator according to the minimax game method, the Wasserstein method and the "penalty gradients" method is described. A digital simulation of the GAN network was performed using a minimax model trained on the MNIST dataset. Conclusions are drawn regarding the possibility of using the developed generative-adversarial neural network in order to expand the bank of training (test) samples.
Key words: neural networks, recognition, generative modeling, training sample, test sample.
Corresponding author: Mark Lvovich Belokopytov, hommer1990@mail.ru
References
1. Malykhina G.F., Merkusheva A.V. Elements of the statistical concept of neural network training and forecasting of its functioning. Nauchnoe priborostroenie [Scientific Instrumentation]. 2005. V.15. №1. P.29-45. (In Russian)
2. Belokopytov M.L., Shlenskikh D.A., Morozov S.V., Sirota S.V. Monitoring of maritime navigation facilities based on aerospace remote sensing data in the microwave range using neural network technologies. Zhurnal radioelektroniki [Journal of Radio Electronics] [online]. 2022. №4. https://doi.org/10.30898/1684-1719.2022.4.2 (In Russian)
3. Okunev S.V. Consideration of ways of forming data sets for training neural networks. Vestnik nauki i obrazovaniya [Bulletin of Science and Education]. 2020. V.3. №2. P.16-19. (In Russian)
4. Ananyin E.V., Andryushchenko M.S. Methods of radar signature research in the development of inconspicuous samples of military equipment. Voprosy oboronnoj texniki. Seriya 16: texnicheskie sredstva protivodejstviya terrorizmu [Issues of defense equipment. Series 16: technical means of countering terrorism]. 2015. №87-88. P.102-108. (In Russian)
5. Tertyshnik V.V., Zinoviev V.V. Methodology for calculating radar portraits of aerodynamic objects. Mezhdunarodnaya nauchno-texnicheskaya konferenciya, posvyashhennaya 15-letiyu kafedry radiotexniki [International Scientific and Technical conference dedicated to the 15th anniversary of the Department of Radio Engineering]. Radio Engineering Department. Saratov. 2004. P.11-24. (In Russian)
6. Filippskikh E.E., Popov A.V., Galkin F.A. Model-oriented methodology for calculating radar portraits (signatures) of metal objects. Zhurnal radioelektroniki [Journal of Radio Electronics] [online]. 2022. №3. https://doi.org/10.30898/1684-1719.2022.3.5 (In Russian)
7. Garcia Gloria Bueno. Obrabotka izobrazheniy s pomoshchyu OpenCV [Image processing using OpenCV]. Moscow, DMK Press, 2015. 387 p. (In Russian).
8. Jonathan Hui. GAN – Wasserstein GAN [web]. Medium. Date of access: 11.11.2022. URL: https://jonathan-hui.medium.com/gan-wasserstein-gan-wgan-gp-6a1a2aa1b490
9. Ishaan Gulrajani, Faruk Ahmed, Martin Arjovsky, Vincent Dumoulin, Aaron Courville. Improved Training of Wasserstein GANs [web]. NIPS 2017. Date of access: 21.11.2022. URL: https://proceedings.neurips.cc/paper/2017/file/892c3b1c6dccd52936e27cbd0ff683d6-Paper.pdf
For citation:
Belokopytov M.L. Using generative-adversarial neural networks to expand training (test) samples. Zhurnal radioelektroniki [Journal of Radio Electronics] [online]. 2023. №2. https://doi.org/10.30898/1684-1719.2023.2.7 (In Russian)