DOI: 10.17587/prin.17.3-13

Development of a Spiking Neural Network for the Quantitative Assessment of Critical Information Infrastructure Security

E. V. Palchevsky¹, Ph.D. (Engineering), Associate Professor of the Department, teelxp@inbox.ru, V. V. Antonov², D. Sc. (Engineering), Head of the Department, antonov.v@bashkortostan.ru, I. I. Khasanov³, Ph. D. (Engineering), Associate Professor of the Department, iikhasanov@fa.ru, V. A. Suvorova², Ph. D. (Engineering), Associate Professor of the Department, milana_da@mail.ru
¹ MIREA — Russian Technological University (RTU MIREA), Moscow,
² Institute of Computer Science, Mathematics and Robotics, Ufa University of Science and Technology, Ufa,
³ Financial University under the Government of the Russian Federation, Moscow

Corresponding author: Evgeny V. Palchevsky, Ph.D., Associate Professor of the Department, Institute of Advanced Technologies and Industrial Programming, MIREA — Russian Technological University (REU MIREA), Moscow, 119454, Russian Federation E-mail: teelxp@inbox.ru

This paper presents the architecture of SecureSpikeNet, a next-generation spiking neural network designed to quantify security risks in the banking sector's critical information infrastructure (CII). Unlike traditional models that rely on computationally intensive gradient-based training, SecureSpikeNet employs a dual-level learning strategy: local training of the hidden layer using extended spike-timing-dependent plasticity (e-STDP) and global optimization of the output layer via the REINFORCE algorithm. Input data are converted into spike trains based on 13 features extracted from NetFlow, IDS, EDR, SIEM and other telemetry sources. The architecture can be deployed on both domestic and international neuromorphic processors (e.g., Intel Loihi 2 and the Russian "Altai" chip) and automatically adapts to an evolving threat landscape. A novel integral risk metric — SecureSpikeScore is introduced, suitable for transmission to regulators and integration into risk-management systems. On a test dataset of 10 000 records, a modified SecureSpikeNet achieved 97.4 % accuracy and an Fl-score of 75.4 %. Compared with CNN-IDS (98.5/98 %) and LSTM-IDS (98/97.5 %), the proposed model delivers comparable detection quality while requiring substantially fewer computational resources and uniquely provides a quantitative threat assessment absent from the reference models.

Keywords: spiking neural network, SecureSpikeNet, critical information infrastructure, quantitative risk assessment, SecureSpikeScore, e-STDP, REINFORCE, neuromorphic processors, streaming telemetry analysis, banking cybersecurity

pp. 3—13

For citation:
Palchevsky E. V., Antonov V. V., Khasanov I. I., Suvorova V. A. Development of a Spiking Neural Network for the Quantitative Assessment of Critical Information Infrastructure Security, Programmnaya Ingeneria, 2026, vol. 17, no. 1, pp. 3—13. DOI: 10.17587/prin.17.3-13. (in Russian).

The article was prepared within the framework of State Assignment No. 10230328002153-1.2.1 of the Financial University under the Government of the Russian Federation.

References:

1. Golembiovskij M. M., Golembiovskaja O. M., Kondrashova E. V. et al. Formalization of the process of assessing the security of critical information infrastructure facilities, Informacionnye sistemy i tehnologii, 2023, no. 4 (138), pp. 92—100 (in Russian).
2. Danilov D. Ju., But N. D. Protection of critical information infrastructure by means of prosecutorial supervision as an important component of national security, Vestnik Universiteta prokuratury Rossijskoj Federacii, 2024, no. 2 (100), pp. 45—52 (in Russian).
3. Alekseeva A. Money transfers without voluntary consent: what Bank of Russia statistics say, Infragreen.ru, 18 May 2025, available at https://infragreen.ru/pierievody-dieniegh-biez-dobrovolnogho-soghlasiia/ (date of access 10.06.2025) (in Russian).
4. Slapnicar S., Vuko T., Cular M., Drascek M. Effectiveness of cybersecurity audit, International Journal of Accounting Information Systems, 2022, vol. 44, article 100548. DOI: 10.1016/j.accinf.2021.100548.
5. Mahboubi A., Luong K., Aboutorab H. et al. Evolving techniques in cyber threat hunting: A systematic review, Journal of Network and Computer Applications, 2024, vol. 232, article 104004. DOI: 10.1016/j.jnca.2024.104004.
6. Zheng Y., Li Z., Xu X., Zhao Q. Dynamic defenses in cyber security: Techniques, methods and challenges, Digital Communications and Networks, 2021, vol. 8, no. 4, pp. 422—435. DOI: 10.1016/j.dcan.2021.07.006.
7. Son K.-S., Song J.-G., Hahm I., Lee J.-W. Quantifying cyber risk: A model for evaluating safety impacts of cyber threats on NPPs, Nuclear Engineering and Technology, 2025, vol. 57, no. 10, article 103675. DOI: 10.1016/j.net.2025.103675.
8. Zadeh A., Lavine B., Zolbanin H. M., Hopkins D. A cybersecurity risk quantification and classification framework for informed risk-mitigation decisions, Decision Analytics Journal, 2023, vol. 9, article 100328. DOI: 10.1016/j.dajour.2023.100328.
9. AlHidaifi S. M., Asghar M. R., Ansari I. S. Towards a Cyber-Resilience Quantification Framework (CRQF) for IT Infrastructure, Computer Networks, 2024, vol. 247, article 110446. DOI: 10.1016/j.comnet.2024.110446.
10. Han J., Kamber M., Pei J. Data Mining: Concepts and Techniques. 3rd ed. Amsterdam, Elsevier, 2011, 744 p.
11. Chandola V., Banerjee A., Kumar V. Anomaly detection: A survey, ACM Computing Surveys, 2009, vol. 41, no. 3, article 15. DOI: 10.1145/1541880.1541882.
12. Garcia S., Grill M., Stiborek J., Zunino A. An empirical comparison of botnet detection methods, Computers & Security, 2014, vol. 45, pp. 100—123. DOI: 10.1016/j.cose.2014.05.011.
13. Dan Y., Poo M. M. Spike timing-dependent plasticity of neural circuits, Neuron, 2004, vol. 44, no. 1, pp. 23—30. DOI: 10.1016/j.neuron.2004.09.007.
14. Poisson S.-D. Researches into the probabilities of judgements in criminal and civil cases / Transl. by O. Sheynin. Berlin, Oscar Sheynin, 2013, 252 p., available at: https://probabilityandfinance.com/sheynin/053_Poisson_en.pdf (date of access 30.06.2025).
15. Williams R. J. Simple statistical gradient-following algorithms for connectionist reinforcement learning, Machine Learning, 1992, vol. 8, pp. 229—256. DOI: 10.1007/BF00992696.
16. Sinha P., Sahu D., Prakash S. et al. A high-performance hybrid LSTM-CNN secure architecture for IoT environments using deep learning, Scientific Reports, 2025, vol. 15, article 9684. DOI: 10.1038/s41598-025-94500-5.
17. Wang Z., Ghaleb F. A., Zainal A. et al. An efficient intrusion detection model based on convolutional spiking neural network, Scientific Reports, 2024, vol. 14, article 7054. DOI: 10.1038/s41598-024-57691-x.
18. Fan S., Paul P., Wu T. et al. On applications of Spiking Neural P Systems, Applied Sciences, 2020, vol. 10, no. 20, article 7011. DOI: 10.3390/app10207011.