Programmable algorithms and neural networks in teaching mathematical disciplines
DOI: 10.23951/2307-6127-2026-2-55-65
The article presents the results of a study of the use of neural networks in teaching mathematical disciplines and a comparison of their capabilities with products based on programmable algorithms: task databases, computer simulators, and mind maps. The questions are considered: can neural networks effectively generate educational content, develop students’ computational skills, provide reliable reference information, and act as tutors in preparation for exams, including the Unified State Exam and the Basic State Exam. The use of artificial intelligence (AI) in educational practice is growing, especially among schoolchildren and students who use neural networks to do homework and cheat on exams. However, experiments show that AI-generated tasks often contain arithmetic errors and logical inconsistencies, and the hallucinations of neural networks make them an unreliable source of information. Comparative tests of ChatGPT, DeepSeek, GigaChat, and Alisa Yandex on Unified State Exam tasks revealed the limited ability of neural networks to solve complex problems of the second part and the inability to solve problems whose condition is specified by a drawing. Computer simulators, task databases and mind maps remain indispensable for the formation of stable mathematical skills: assimilation of the logical structure of disciplines and memorization of mathematical facts. Important advantages of these tools: generation of an unlimited number of reliable exercises for the development of mathematical skills; interface providing instant feedback; tracking of students’ progress through a database. It has been shown that neural networks can be useful for generating educational materials and analyzing large unrefined data, identifying non-trivial recommendations for improving the educational process. However, they are not able to monitor the student’s activity, identify individual gaps and adjust the educational trajectory, which makes them a weak substitute for a teacher or tutor. The optimal strategy for integrating AI into the educational process is to combine the capabilities of neural networks with proven tools based on programmable algorithms. The teacher, adjusting and checking the results of the AI, can increase the effectiveness of the educational process, while maintaining the reliability and quality of the formation of knowledge and skills.
Keywords: neural networks, artificial intelligence, programmable algorithms, teaching mathematics, computational skills, generation of educational content, Unified State Exam, Basic State Exam, mind maps, computer simulators
References:
1. Stychinskiy M.S. Primeneniye iskusstvennykh kognitivnykh sistem v obrazovanii i nauke [Application of artificial cognitive systems in education and science]. Istoriya, 2024, vol. 15, no. 12-2 (146) (in Russian). DOI: 10.18254/S207987840032014-4
2. Uzhe kazhdyy chetvyortyy roditel’ uchenika 5–6 klassov znayet, chto rebyonok vypolnyayet “domashku” s II [Every fourth parent of a 5th-6th grade student already knows that their child is doing homework with AI] (in Russian). https://www.superjob.ru/research/articles/115357/uzhe-kazhdyj-chetvertyj-roditel-uchenika-5/ (accessed 5 September 2025).
3. Belkin D. They Were Every Student’s Worst Nightmare. Now Blue Books Are Back. The Wall Street Journal, 28 January 2023 g. URL: https://www.wsj.com/business/chatgpt-ai-cheating-college-blue-books-5e3014a6 (accessed 5 September 2025).
4. Mehlig B. Machine learning with neural networks: an introduction for scientists and engineers. Cambridge, Cambridge University Press, 2022. 250 p.
5. Minakov A.I. Iskusstvennyy intellekt i neyroseti v obrazovanii: uchebnik [Artificial Intelligence and Neural Networks in Education: textbook]. Moscow, Direkt-Media Publ., 2024. 164 p. (in Russian)
6. Sdam GIA [I will pass the State Final Attestation] (in Russian). URL: // https://sdamgia.ru/ (accessed 5 September 2025).
7. Onlayn podgotovka po matematike i fizike [Online preparation in mathematics and physics] (in Russian). URL: https://www.workingmemory.ru/ (accessed 5 September 2025).
8. Pravilo tryokh klikov [The Three Click Rule] (in Russian). URL: https://ru.wikipedia.org/wiki/Правило_трех_кликов (accessed 5 September 2025).
9. B’yuzen T. Intellekt-karty. Polnoye rukovodstvo po moshchnomu instrumentu myshleniya [Mind Maps: A Complete Guide to a Powerful Thinking Tool]. Moscow, Mann, Ivanov i Ferber Publ., 2021. 208 p. (in Russian).
10. Intellekt karta po matematike [Mind map in mathematics] (in Russian). URL: https://www.workingmemory.ru/pdf/mind_map_on_math_ege.pdf (accessed 5 September 2025).
11. Intellekt karta po fizike [Mind map on physics] (in Russian). URL: https://www.workingmemory.ru/pdf/mind_map_on_phys_ege.pdf (accessed 5 September 2025).
12. Krutetskiy V.A. Psikhologiya matematicheskikh sposobnostey shkol’nikov [Psychology of mathematical abilities of schoolchildren]. Ed. N.I. Chuprikova. Moscow, Institute of practical psychology; Voronezh, MODEK Publ., 1998. 416 p. (in Russian).
13. Dehn Milton J. Long-Term Memory Problems in Children and Adolescents: Assessment, Intervention, and Effective Instruction. John Wiley & Sons, Inc., 2010. 408 p.
14. Prokof’yev A.A., Sokolova T.V. Matematika. Profil’nyy uroven’. Yedinyy gosudarstvennyy ekzamen. Gotovimsya k itogovoy attestatsii: uchebnoye posobiye [Mathematics. Profile level. Unified state exam. Preparing for final certification: textbook]. Moscow, Intellekt-Tsentr Publ., 2025. 248 p. (in Russian).
15. EGE-2025: rezul’taty sopostavimy s proshlym godom, no po matematike – proryv [Unified State Exam 2025: Results Comparable to Last Year, but a Breakthrough in Mathematics] (in Russian). URL: https://rg.ru/2025/06/08/priglasili-na-ball.html (accessed 5 September 2025).
16. Sootvetstviye mezhdu pervichnymi ballami i testovymi ballami EGE 2025 [Correspondence between primary scores and test scores of the Unified State Exam 2025] (in Russian). URL: https://obrnadzor.gov.ru/wp-content/uploads/2025/05/tabliczy_sootvetstvij_2025.pdf#page=6.00 (accessed 5 September 2025).
17. DeepSeek v deystvii [DeepSeek in Action]. Translation from Chinese V.S. Yatsenkova. Moscow, DMK Press Publ., 2025. 404 p. (in Russian).
( 611.3 kB ) 
( 599.09 kB )
Issue: 2, 2026
Series of issue: Issue 2
Rubric: THEORY AND METHODOLOGY OF TEACHING AND EDUCATION
Pages: 55 — 65
Downloads: 20