Neural Networks Optimization: Methods and Their Comparison Based off Text Intellectual Analysis
- Authors: Torkunova J.V.1,2, Milovanov D.V.1
-
Affiliations:
- Kazan State Power Engineering University
- Sochi State University
- Issue: Vol 13, No 4 (2023)
- Pages: 142-158
- Section: Articles
- Published: 30.12.2023
- URL: https://journals.rcsi.science/2328-1391/article/view/348634
- DOI: https://doi.org/10.12731/2227-930X-2023-13-4-142-158
- EDN: https://elibrary.ru/SFIPKW
- ID: 348634
Cite item
Full Text
Abstract
The research resulted in the development of software that implements various algorithms of neural networks optimization, which allowed to carry out their comparative analysis in terms of optimization quality. The article takes a detailed look at artificial neural networks and methods of their optimization: quantization, overcutting, distillation, Tucker’s dissolution. Algorithms and optimization tools of neural networks were explained, as well as comparative analysis of different methods was conducted with their advantages and disadvantages listed. Calculation values were given as well as recommendations on how to execute each method. Optimization is studied by text classification performance: peculiarities were removed, models were chosen and taught, parameters were adjusted. The set task was completed with the use of the following technologies: Python programming language, Pytorch framework and Jupyter Notebook developing environment. The results that were acquired can be used to reduce the demand on computing power while preserving the same level of detection and classification abilities.
About the authors
Julia V. Torkunova
Kazan State Power Engineering University; Sochi State University
Author for correspondence.
Email: torkynova@mail.ru
ORCID iD: 0000-0001-7642-6663
SPIN-code: 7422-4238
Professor of the Department of Information Technologies and Intelligent Systems, Doctor of Pedagogical Sciences
Russian Federation, 51, Krasnoselskaya Str., Kazan, Republic of Tatarstan, 420066, Russian Federation; 94, Plastunskaya Str., Sochi, Krasnodar region, 354000, Russian Federation
Danila V. Milovanov
Kazan State Power Engineering University
Email: studydmk@gmail.com
Magister
Russian Federation, 51, Krasnoselskaya Str., Kazan, Republic of Tatarstan, 420066, Russian Federation
References
- Avetisyan T. V., L’vovich Ya. E. International Journal of Advanced Studies, 2023, vol. 13, no. 1, pp. 102-114. https://doi.org/10.12731/2227-930X-2023-13-1-102-114
- Akzholov R.K., Veriga A.V. Vestnik nauki, 2020, no. 3 (24), pp. 66-68.
- Akhmetzyanova K.R., Tur. A.I., Kokoulin A.N. Vestnik Permskogo natsional’nogo issledovatel’skogo politekhnicheskogo universiteta. Elektrotekhnika, informatsionnye tekhnologii, sistemy upravleniya, 2020, no. 36, pp. 117-130. https://doi.org/10.15593/2224-9397/2020.4.07
- Kashirina I. L., Demchenko M. V. Vestnik VGU, seriya: Sistemnyy analiz i informatsionnye tekhnologii, 2018, no. 4, pp. 123-132.
- Kopyrin A. S., Makarova I. L. Programmnye sistemy i vychislitel’nye metody, 2020, no. 3, pp. 40-50. https://doi.org/10.7256/2454-0714.2020.3.33958
- Osovskiy S. Neyronnye seti dlya obrabotki informatsii [Neural networks for information processing]. Moscow: Hot Line. Telecom. 2019, 448 p.
- Romanov D.E. Inzhenernyy vestnik Dona, 2009, no. 3, pp. 19-24.
- Sozykin A.V. Vestnik YuUrGU. Seriya: Vychislitel’naya matematika i informatika, 2017, no. 3 (6), pp. 28-59.
- Torkunova Yu.V., Korosteleva D.M., Krivonogova A.E. Sovremennoe pedagogicheskoe obrazovanie, 2020, no. 5, pp. 107-110.
- Cherkasova I.S. E-Scio, 2022. https://e-scio.ru/wp-content/uploads/2022/03/%D0%A7%D0%B5%D1%80%D0%BA%D0%B0%D1%81%D0%BE%D0%B2%D0%B0-%D0%98.-%D0%A1.pdf
- Yashchenko A.V., Belikov A.V., Peterson M.V. Nauchno-tekhnicheskiy vestnik informatsionnykh tekhnologiy, mekhaniki i optiki, 2020, no. 3, pp. 402-409.
- A White Paper on Neural Network Quantization. https://doi.org/10.48550/arXiv.2106.08295
- Distilling Task-Specific Knowledge from BERT into Simple Neural Networks. https://doi.org/10.48550/arXiv.1903.12136
- Majid Janzamin, Rong Ge, Jean Kossaifi and Anima Anandkumar. Spectral Learning on Matrices and Tensors. Foundations and Trends R in Machine Learning, 2019, vol. 12, no. 5-6, pp. 393–536. https://doi.org/10.1561/2200000057
- Tensor Networks for Latent Variable Analysis. Part I: Algorithms for Tensor Train Decomposition. https://arxiv.org/pdf/1609.09230.pdf
Supplementary files



