Neuroeducational Environment for Acquisition of Competencies in the Field of End-To-End Digital Technologies (Neurotechnology) in the Conditions of Digital Transformation

Purpose of the study. The purpose of the study is to develop proposals for the formation of a neuroeducational environment that ensures that students acquire the competencies necessary to solve the problem of introducing end-to-end digital technologies within the framework of the digital economy program. The analysis showed that improving the training of specialists in the field of neurotechnology and artificial intelligence in the digital economy is an urgent and demanded task. The paper discusses an approach to solve the problem of improving the training of specialists in the field of neurotechnologies, taking into account the requirements of the federal project “Personnel for the digital economy” by creating a technological platform for a neuroeducational environment that ensures the acquisition of the necessary competencies by students.

Materials and methods of research. In the process of carrying out the research, theoretical provisions in the field of the theory of neural networks were developed, a new class of neurons and neural networks was proposed, which are close in functions to biological neural networks and are called selective. The technology of training and application of selective neural networks has been developed. The advantage of selective technologies in comparison with classical neural networks is shown. Hardware models of neural networks, which are part of the neuroeducational system, have been developed and used in the educational process. A general methodology for teaching neural network technology as one of the end-to-end technologies of the digital economy has been developed, as well as a methodology for using neural networks in solving economic problems in the context of digital transformation.

Results. A technological platform for a neuroeducational environment has been developed, including software, hardware models of classical neurons and perceptrons (McCulloch-Pitts), as well as neurons and perceptrons of a new class, called selective. A software tool for teaching standard and selective neurotechnologies was developed and proposed, for which a certificate of state registration of a computer program was obtained. For the assimilation of theoretical material and the acquisition of practical skills, methodological materials, author’s practical tasks and author’s laboratory works have been developed, which are part of the technological platform. The approaches proposed in the article can be used in organizing the study of the theory of neural networks and methods of applied application of neurotechnologies in solving the problems of introducing end-to-end digital technologies within the framework of the digital economy program. The development results are confirmed by 4 patents for inventions. In order to more effectively master the theoretical provisions and features of the practical application of neurotechnologies, the main attention is paid to the physical meaning and presentation of the processes occurring during the functioning of a neural network in the form of formal descriptions that provide a more effective assimilation of the foundations of the theory of neural networks and neurotechnologies using existing standard neural network architectures, as well as architectures built on the basis of selective neural networks.

Conclusion. The architecture and components of the technological platform of the neuroeducational environment based on neuroeducational complexes have been developed. A general methodological approach has been developed for teaching the basics of neurotechnology based on standard and selective neural networks and the peculiarities of their application in the framework of the digital economy program. A methodology for teaching the basics of neurotechnology based on standard and selective neural networks has been developed, which includes the mathematical theory of standard and selective neural networks, a description of the learning process for standard neural networks based on McCulloch-Pitts neurons, as well as selective neural networks based on selective neurons.

1. Rasporyazheniye Pravitel’stva RF ot 28.07.2017 N 1632-r ob utverzhdenii programmy «Tsifrovaya ekonomika Rossiyskoy Federatsii» = Order of the Government of the Russian Federation of July 28, 2017 N 1632-r on the approval of the program «Digital Economy of the Russian Federation» [Internet]. Available from: http://base.garant.ru/71734878/ (cited 14.10.2020). (In Russ.)

2. Natsional’naya programma «Tsifrovaya ekonomika Rossiyskoy Federatsii», utverzhdena protokolom zasedaniya prezidiuma Soveta pri Prezidente Rossiyskoy Federatsii po strategicheskomu razvitiyu i natsional’nym proyektam ot 4 iyunya 2019 g. № 7 = The national program «Digital Economy of the Russian Federation», approved by the minutes of the meeting of the Presidium of the Council under the President of the Russian Federation for Strategic Development and National Projects of June 4, 2019 No. 7. [Internet]. Available from: https://digital.gov.ru/ru/activity/directions/858/ (cited 14.10.2020). (In Russ.)

3. Sayt Ministerstva tsifrovogo razvitiya, svyazi i massovykh kommunikatsiy. Rossiyskoy Federatsii. Skvoznyye tsifrovyye tekhnologii = Website of the Ministry of Digital Development, Communications and Mass Media. Russian Federation. End-toend digital technologies [Internet]. Available from: https://digital.gov.ru/ru/activity/directions/878/(cited 14.10.2020). (In Russ.)

4. Publichnyy analiticheskiy doklad po napravleniyu «neyrotekhnologii» = Public analytical report in the direction of «neurotechnology» [Internet]. Available from:https://reestr.extech.ru/docs/analytic/reports/neuroscience.pdf (cited 14.10.2020). (In Russ.)

5. Dorozhnaya karta razvitiya «skvoznoy» tsifrovoy tekhnologii «neyrotekhnologii i iskusstvennyy intellekt» 10 oktyabrya 2019 = Roadmap for the development of «end-to-end» digital technology «neurotechnology and artificial intelligence» October 10, 2019 [Internet] Available from: https://digital.gov.ru/ru/documents/6658/. (cited 14.10.2020). (In Russ.)

6. Plan meropriyatiy «dorozhnaya karta» «Neyronet» Natsional’noy tekhnologicheskoy initsiativy = Action plan «road map» «Neuronet» of the National Technology Initiative [Internet]. Available from: https://nti2035.ru/markets/neuronet(cited 14.10.2020). (In Russ.)

7. Prikaz Minekonomrazvitiya ot 24.01.2020g. № 41 «Ob utverzhdenii metodik rascheta pokazateley proyekta «Kadry dlya tsifrovoy ekonomiki» dlya natsional’noy programmy «Tsifrovaya ekonomika» = Order of the Ministry of Economic Development of 01/24/2020. No. 41 «On the approval of methods for calculating the indicators of the project» Personnel for the digital economy «for the national program» Digital economy «[Internet]. Available from: https://rulaws.ru/acts/Prikaz-Minekonomrazvitiya-Rossiiot-24.01.2020-N-41/ (cited 14.10.2020). (In Russ.)

8. Mazurov M. Ye. Impul’snyy neyron, blizkiy k real’nomu. Patent na izobreteniye № 2598298. 09.02.2015 = Impulse neuron, close to real. Patent for invention No. 2598298. 09.02.2015. (In Russ.)

9. Mazurov M. Ye. Neyron, modeliruyushchiy svoystva real’nogo neyrona. Patent na izobreteniye № 2597495. 07.11.2014 = Neuron modeling the properties of a real neuron. Patent for invention No. 2597495. 07.11.2014. (In Russ.)

10. Mazurov M. Ye. Odnosloynyy pertseptron na osnove izbiratel’nykh neyronov. Patent na izobreteniye № 2597497. 13.01.2015 = Single-layer perceptron based on selective neurons. Patent for invention No. 2597497. 13.01.2015. (In Russ.)

11. Mazurov M. Ye. Odnosloynyy pertseptron, modeliruyushchiy svoystva real’nogo pertseptrona. Patent na izobreteniye № 2597496. 24.02.2015 = Single-layer perceptron, modeling the properties of a real perceptron. Patent for invention No. 2597496. 24.02.2015. (In Russ.)

12. Pertseptrony = Perceptrons [Internet]. Available from: https://neuralnet.info/chapter/perseptrony/ (cited 14.10.2020). (In Russ.)

13. Novikoff A. B. On convergence proofs on perceptrons. Symposium on the Mathematical Theory of Automata. Polytechnic Institute of Brooklyn. 1962. 12: 615-622.

14. Mazurov M. Ye. Accuracy of image recognition in selective neural networks. Neyrokomp’yutery: razrabotka, primeneniye = Neurocomputers: development, application. 2020. (In Russ.)

15. Mazurov M. Ye. Nonlinear dynamics and synchronization of neural ensembles in the formation of attention. Izvestiya RAN. Seriya fizicheskaya = Izvestiya RAN. Physical series. 2020; 84; 3: 451–456. (In Russ.)

16. Mazurov M. Ye. Svidetel’stvo o Gosudarstvennoy registratsii «Programma rascheta vesovykh koeffitsiyentov pertseptrona s pomoshch’yu izbiratel’nogo metoda Monte-Karlo». № 2019617233. 04.06.2019 = Certificate of State Registration «Program for calculating the weight coefficients of the perceptron using the selective Monte Carlo method.» No.2019617233 dated 06/04/2019. (In Russ.)

17. Galushkin A. I. Neyronnyye seti: osnovy teorii = Neural Networks: Theory Foundations. Moscow: Hot line – telecom; 2010. 496 p. (In Russ.)

18. Mazurov M. Ye. Teaching selective binary neural networks without mathematics and without a teacher using self-organization. XVIII Vserossiyskaya nauchnaya konferentsiya «Neyrokomp’yutery i ikh primeneniye» = XVIII All-Russian scientific conference «Neurocomputers and their application». 2020: 27–28. (In Russ.)

19. Kirichenko A.A. «Neyropakety – sovremennyy intellektual’nyy instrument issledovatelya». Setevoye elektronnoye izdaniye uchebnogo posobiya =»Neuropackages are a modern intellectual tool for a researcher.» Online electronic publication of the textbook. 2016. 297p. [Internet] Available from: https://publications.hse.ru/books/91277065 (cited 14.10.2020). (In Russ.)

20. Deductor «O sisteme» = Deductor «About the system» [Internet] Available from: https://basegroup.ru/deductor/description (cited 14.10.2020). (In Russ.)

21. Prokopenko YU.N. Sistemy podderzhki prinyatiya resheniy = Decision support systems. Nizhny Novgorod: NNGASU; 2017. 188 p. (In Russ.)