student
The article discusses the solution of the object classification problem based on the use of impulse neural networks (SNN). These networks effectively simulate the biological principles of the brain by using discrete impulses. This approach opens up new possibilities for solving problems where the speed and accuracy of information processing, as well as energy efficiency, are important. One of the most popular and simplest models of neurons in pulsed neural networks is the LeakyIntegrate-and-Fire (LIF) model, which describes the dynamics of a neuron's membrane potential as a process of leakage and integration of input current. This process reflects the biological behavior of neurons, where information is accumulated and transmitted through impulses. The source data in pulsed neural networks is encoded as time sequences of pulses, which allows efficient processing of information that changes over time. The input neurons produce impulses that correspond to the time points reflected in the input sequence. When the membrane potential of a neuron reaches a set threshold, it is activated, which leads to the generation of a pulse, after which the potential of the neuron is reset to zero, and the process repeats. The article provides examples of using pulsed neural networks to solve two classical classification problems: irises and wheat varieties, which are traditionally used by researchers to evaluate the effectiveness of the algorithms being developed. The paper considers hyperparameters of an impulse neural network that affect classification results, including the number of presynaptic neurons, the weights of connections between neurons, the threshold of neuron activation, membrane resistance, resting potential, and other hyperparameters.
ARTIFICIAL INTELLIGENCE, IMPULSE NEURAL NETWORKS, SPIKE, CLASSIFICATION
1. Stepanov, S. U., Sidorenko, A. U., Kolbina, O. N., Martyn, I. A. Information Technologies and Systems: Management, Economics, Transport, Law, 2024, No. 3(51), 157-162. EDN ECBDLQ.
2. Lazarev, E. A. Bulletin of Science, 2023, Vol. 5, No. 12-1(69), 412-415. EDN BVXPYI.
3. Khol, E. V., Sharonova, K. S. Modern Innovations, Systems and Technologies, 2023, Vol. 3, No. 4, 311-321. DOI:https://doi.org/10.47813/2782-2818-2023-3-4-0311-0321. EDN WMFPMH.
4. Pankrashov, A. S., Okladnikova, S. V. Engineering and Construction Bulletin of the Caspian Region, 2025, No. 4(54), 63-69. DOI:https://doi.org/10.52684/2312-3702-2025-54-4-63-69. EDN KQHLRC.
5. Emaletdinova, L. U., Kabirova, A. N. News of Higher Educational Institutions. Aviation Technology, 2019, No. 2, 40-48. EDN FEXHKC.
6. Nazipov, R. S. Modern Science: Relevant Problems of Theory and Practice. Series: Natural and Technical Sciences, 2024, No. 9, 101-103. DOI:https://doi.org/10.37882/2223-2966.2024.9.27. EDN TFINYY.
7. Lesnikov, S. V. Govor: Almanac, 2025, No. 6, 686-693. EDN UOMKNB.
8. Cherpakov, I. V. EFO: Economics. Finance. Society, 2025, No. 3(15), 110-121. DOI:https://doi.org/10.24412/2782-4845-2025-15-110-121. EDN ISTUPY.
9. Khudaibiredeva, G. B., Kozukhov, D. A., Pimenova, A. A. Theory and Practice of Modern Science, 2025, No. 8(122), 98-105. EDN UHQCRK.
10. Aglamzyanova, G. N., Gumerova, L. Z., Garipova, R. F. Socio Economic and Technical Systems: Research, Design, Optimization, 2023, No. 1(93), 6-12. EDN GSOWTQ.
11. Antonov, A. M., Shchanikov, S. A. Methods and Devices of Information Transmission and Processing, 2023, No. 25, 43-52. EDN CQURGH.
12. Kiselev, M. V., Urosov, A. M., Ivanitsky, A. U. Computer Research and Modeling, 2025, Vol. 17, No. 3, 389-400. DOI:https://doi.org/10.20537/2076-7633-2025-17-3-389-400. EDN ISXQIQ.
13. Matveev, N. I. Innovations. Science. Education, 2021, No. 45, 416-421. EDN LONVHM. DOI: https://doi.org/10.15405/epsbs.2021.07.02.50
14. Palychevsky, E. V., Antonov, V. V., Khasanov, I. I., Suvorova, V. A. Software Engineering, 2026, Vol. 17, No. 1, 3-13. DOI:https://doi.org/10.17587/prin.17.3-13. EDN HLACFS.
15. First steps in spiking neural networks. – URL: https://habr.com/ru/articles/746762/ (accessed 20.11.2025).
