S.V. Smirnov1

1 Trapeznikov Institute of Management Problems of RAS (Moscow, Russia)
1 sapr2006@bk.ru

The work is devoted to the use of neural networks, which successfully solve the most urgent problems facing modern society in various fields. This is due to the versatility of the methodological approach in solving various problems (scientific, technical and commercial): pattern recognition, object identification, clustering, information compression, visualization of multidimensional data, development of expert systems, writing technical and commercial reviews, design of complex systems, etc. Neural networks are already used in such industries as: medicine, construction, aircraft and rocket science, cinematography, forensics, visual arts, management, machine and machine tools, trade, etc. (the list of applications is regularly increasing).

The purpose of the work is to consider certain examples of the practical application of neural networks, to indicate the selective spheres in which they function to solve various problems, as well as to consider a simplified mechanism of their work.

As a result of the work, the application and functioning of the following neural networks was considered: ChatGPT in the Bing Internet search engine of Microsoft, YandexGPT in the Internet search engine of the same name of the Russian company Yandex, MENNDL in the automotive industry of the American corporation General Motors, MDSS (medical decision support system), introduced by the government Moscow to medical institutions in the capital, AlphaFold (protein structure prediction) developed by the British company Google DeepMind.

The neural network was also considered as part of the Russian intelligent robotic system for copying left-handed movements (LevshAI) for remote neurosurgical endovascular operations. In the field of representation of the graphic model, the Neuralangelo neural network, developed by the American company NVIDIA Research to improve 3D reconstructions, and the ModelScope text-to-video conversion generator, created by one of the divisions of the Chinese Internet giant Alibaba, were considered. In addition, the use of the PARSIV neural network, developed by the Moscow Department of Information Technologies (DIT) for use in the Ministry of Internal Affairs, was considered.

The practical significance lies in the fact that the considered areas of application of specific (private) neural networks: Internet, video production, automotive, medicine, architecture, law enforcement system, are not well studied in various review publications. This publication is able to contribute to clarifying some of the poorly detailed reviews on the use of private neuroscience.

It should be noted that the examples of the use of neural networks described in this work are of a scientific and practical nature and contribute to the popularization of the promotion of neurotechnologies in Russia, since quite important systems are consecrated with the introduction of them into the mechanism of their functioning. The future of neural networks seems to be a promising and growing direction in the field of artificial intelligence, which will continue to develop and improve in the coming years.

Smirnov S.V. Some examples of practical application of neural networks: spheres and mechanism of work. Science Intensive Technologies. 2023. V. 24. № 7. P. 67−78. DOI: https://doi.org/10.18127/ j19998465-202307-08 (in Russian)

1. Nejronnaya set'. URL: https://ru.wikipedia.org/wiki/Nejronnaya_set' (data obrashcheniya: 5.06.2023) (in Russian).
2. Buharin S.V., Navoev V.V. Metody teorii nejronnyh setej v ekspertize tekhnicheskih i ekonomicheskih ob"ektov: monografiya. Voronezh: Nauchnaya kniga. 2015. 255 s. (in Russian).
3. Mak-Kallok U. URL: https://ru.wikipedia.org/wiki/Mak-Kallok,_Uorren (data obrashcheniya: 5.06.2023) (in Russian).
4. Perceptron. URL: https://ru.wikipedia.org/wiki/Perceptron (data obrashcheniya: 5.06.2023) (in Russian).
5. Borovskaya E.V., Davydova N.A. Osnovy iskusstvennogo intellekta: Ucheb. posobie. M.: BINOM. Laboratoriya znanij. 2010. 127 s. (in Russian).
6. Terekhov S.A. Lekcii po teorii i prilozheniyam iskusstvennyh nejronnyh setej. URL: https://studylib.ru/doc/807187/lekcii-po-teorii-i-prilozheniyam-iskusstvennyh-nejronnyh-setej (data obrashcheniya: 5.06.2023) (in Russian).
7. Gorban' A.N. Obuchenie nejronnyh setej. M.: SP «ParaGraf». 1990. 159 s. (in Russian).
8. Real-Life and Business Applications of Neural Networks. URL: https://www.smartsheet.com/neural-network-applications (data obrashcheniya: 17.06.2023) (in Russian).
9. Differences between ChatGPT Vs Search Engine. URL: https://beingassistant.com/differences-between-chatgpt-vs-search-engine/ (data obrashcheniya 17.06.2023) (in Russian).
10. Chto umeet II-poiskovik Bing ot Microsoft: testiruem nejroset'. URL: https://journal.tinkoff.ru/bing-ai/ (data obrashcheniya: 17.06.2023) (in Russian).
11. Microsoft dobavila ChatGPT v poisk Bing. Chto eto znachit dlya rynka? URL: https://skillbox.ru/media/code/microsoft-dobavila-chatgpt-v-poisk-bing-chto-eto-znachit-dlya-rynka/ (data obrashcheniya: 17.06.2023) (in Russian).
12. Nejroset' Yandeksa pishet pis'ma i sochinyaet skazki: glavnoe o YandexGPT URL: https://trends.rbc.ru/trends/industry/ 6465e4d39a79474b00a85e06 (data obrashcheniya: 12.06.2023) (in Russian).
13. Najdetsya chto: «Yandeks» zapustit dlya poiskovika sverhsovremennuyu nejroset'. URL: https://iz.ru/1462961/irina-tcyruleva/naidetsia- chto-iandeks-zapustit-dlia-poiskovika-sverkhsovremennuiu-neiroset (data obrashcheniya: 12.06.2023) (in Russian).
14. Chto takoe generativnaya nejroset' YaLM ot Yandeks. URL: https://blog.miralinks.ru/2021/06/15/yandex-yalm/ (data obrashcheniya: 12.06.2023) (in Russian).
15. General Motors licenses AI software developed at Oak Ridg. URL: https://aibusiness.com/verticals/general-motors-licenses-ai-software-developed-at-oak-ridge (data obrashcheniya: 14.06.2023) (in Russian).
16. General Motors Obtains License For MENNDL Artificial Intelligence Software. URL: https://gmauthority.com/blog/2021/04/general-motors-obtains-license-for-menndl-artificial-intelligence-software/ (data obrashcheniya: 14.06.2023) (in Russian).
17. Platforma prognoznoj analitiki Webiomed. URL: https://webiomed.ru/ (data obrashcheniya: 20.06.2023) (in Russian).
18. Vrachi v stolichnyh poliklinikah v 46% sluchayah vybirayut predlozhennyj nejroset'yu diagnoz. URL: https://icmos.ru/news/vrachi-v-stolichnykh-poliklinikakh-v-46-sluchayakh-vybirayut-predlozhennyy-neyrosetyu-diagnoz (data obrashcheniya: 20.06.2023) (in Russian).
19. Stavit' diagnoz s pomoshch'yu iskusstvennogo intellekta v poliklinikah stali eshche tochnee. URL: https://www.mos.ru/news/item/ 93154073/ (data obrashcheniya: 20.06.2023) (in Russian).
20. AlphaFold: Protein Structure Database. URL: https://alphafold.ebi.ac.uk/ (data obrashcheniya: 21.06.2023).
21. AlphaFold. URL: https://ru.wikipedia.org/wiki/AlphaFold (data obrashcheniya: 21.06.2023) (in Russian).
22. Highly accurate protein structure prediction with AlphaFold. URL: https://www.nature.com/articles/s41586-021-03819-2 (data obrashcheniya: 21.06.2023) (in Russian).
23. Robota-hirurga osnastili tekhnologiej peredachi taktil'nyh oshchushchenij. URL: https://misis.ru/science/achievements/2023-06/8593/ (data obrashcheniya: 27.06.2023) (in Russian).
24. Uchenye sozdali tekhnologiyu peredachi taktil'nyh oshchushchenij dlya robotov-hirurgov. URL: http://avtprom.ru/news/2023/06/26/ uchenye-sozdali-tekhnolo (data obrashcheniya: 28.06.2023) (in Russian).
25. LEVShAI ECOSYSTEM. URL: https://neurosputnik.com/levshai/ (data obrashcheniya: 27.06.2023).
26. Digital Renaissance: NVIDIA Neuralangelo Research Reconstructs 3D Scenes. URL: https://blogs.nvidia.com/blog/2023/06/ 01/neuralangelo-ai-research-3d-reconstruction/ (data obrashcheniya: 25.06.2023) (in Russian).
27. Neuralangelo ot NVIDIA – II generiruyushchee iz 2D-video v 3D-ob'ekty. URL: https://vc.ru/future/715152-neuralangelo-ot-nvidia-ii-generiruyushchee-iz-2d-video-v-3d-obekty (data obrashcheniya: 25.06.2023) (in Russian).
28. Nvidia Research Creates Neuralangelo AI Model For 3D Reconstructions. URL: https://www.beyondgames.biz/37679/nvidia-research-creates-neuralangelo-ai-model-for-3d-reconstructions/ (data obrashcheniya: 25.06.2023) (in Russian).
29. The Best Image-to-3D AI to date: Neuralangelo. URL: https://www.louisbouchard.ai/neuralangelo/ (data obrashcheniya: 25.06.2023) (in Russian).
30. ModelScope Text-To-Video. URL: https://theresanaiforthat.com/ai/modelscope-text-to-video/ (data obrashcheniya: 28.06.2023) (in Russian).
31. Text-to-video-synthesis Model in Open Domain. URL: https://modelscope.cn/models/damo/text-to-video-synthesis/summary (data obrashcheniya: 28.06.2023) (in Russian).
32. Diffuzionnaya nejroset' ModelScope text2video 1.7B – sozdaem video po tekstovomu opisaniyu u sebya doma. URL: https://habr.com/ ru/articles/724284/ (data obrashcheniya: 28.06.2023) (in Russian).
33. Raspoznavanie stolichnosti. URL: https://www.kommersant.ru/doc/5029702 (data obrashcheniya: 26.06.2023) (in Russian).
34. Podsistema avtomaticheskoj registracii scenariev indeksirovaniya videoinformacii (PARSIV). URL:https://www.tadviser.ru/index.php/ Produkt:Podsistema_avtomaticheskoj_registracii_scenariev_indeksirovaniya_videoinformacii_(PARSIV) (data obrashcheniya: 26.06.2023) (in Russian).
35. Meriya Moskvy dorabotaet sistemu raspoznavaniya lic special'no dlya MVD. URL: https://www.vedomosti.ru/technology/ news/2021/09/02/884804-sistemu-raspoznavaniya-lits (data obrashcheniya: 26.06.2023) (in Russian).
36. Lovis', prestupnik, v nejroseti. URL: https://ai-news.ru/2022/06/lovis_prestupnik_v_nejroseti.html (data obrashcheniya: 26.06.2023) (in Russian).
37. Barskij A.B. Logicheskie nejronnye seti: Ucheb. posobie. M.: Internet-Universitet Informacionnyh Tekhnologij; BINOM. Laboratoriya znanij. 2007. 352 s. (in Russian).
38. Aksenov S.V., Novosel'cev V.B. Organizaciya i ispol'zovanie nejronnyh setej (metody i tekhnologii). Pod obshch. red. V.B. Novosel'ceva. Tomsk: Izd-vo NTL. 2006. 128 s. (in Russian).
39. Yasnickij L.N. Intellektual'nye sistemy: Uchebnik. M.: Laboratoriya znanij. 2016. 221 s. (in Russian)
40. Golovinskij P.A., Surovcev I.S. Intellektual'nye informacionnye sistemy: teoreticheskie osnovy i prilozheniya. Voronezh: Cifrovaya poligrafiya. 2015. 204 s. (in Russian).
41. Krug P.G. Nejronnye seti i nejrokomp'yutery: Ucheb. posobie po kursu «Mikroprocessory». M.: Izdatel'stvo MEI. 2002. 176 s. (in Russian).
42. Belousov V.V., Druzhinina O.V., Korepanov E.R., Makarenkova I.V., Maksimova V.V. Primenenie nejronnyh setej dlya resheniya zadach klassifikacii pri vyyavlenii neispravnostej transportnyh system. Nejrokomp'yutery: razrabotka, primenenie. 2022. T. 24. № 4. S. 18–27. DOI: https://doi.org/ 10.18127/j19998554-202204-02 (in Russian).
43. Vlasov A.I., Gyul'malieva S.E., Liber YU.S., Abdulkader S. Primenenie sistem iskusstvennogo intellekta v elektrokardiografii. Nejrokomp'yutery: razrabotka, primenenie. 2022. T. 24. № 4. S. 36–52. DOI: https://doi.org/10.18127/j19998554-202201-04 (in Russian)