Digital twins in intrusion detection systems based on deep learning

O.O. Letychevskyi, S.O. Yevdokymov

Abstract


This work aims to improve the accuracy of attack detection in software and hardware systems by utilizing a digital twin in the form of an algebraic model within intrusion detection systems (IDSs) based on deep learning neural networks (DNNs). This approach addresses the shortcomings of training and dataset imperfections that lead to numerous false positives, undetected intrusions, and weak resistance against adversarial attacks. We propose an IDS architecture that combines deep learning neural networks with an algebraic model at the required level of abstraction. This composition provides high detection accuracy and enables continuous self-learning of the IDS based on model operation and data acquisition, including zero-day attacks. Two examples demonstrate the application of this approach: detecting attacks in the binary code of a software system and in a programmable integrated circuit.

Problems in programming 2025; 2: 20-27


Keywords


deep learning; neural network; neurosymbolic approach; digital twins; behavior algebra

References


A. Pinto, L.-C. Herrera, Y. Donoso, and J. A. Gutiérrez, "Survey on intrusion detection systems based on machine learning techniques for the protection of critical infrastructure", Sensors, vol. 23, no. 5, p. 415, Mar. 2023.

D.-M. Ngo, A. Temko, C. C. Murphy, and E. Popovici, "FPGA hardware acceleration framework for anomaly-based intrusion detection system in IoT", in Proc. 31st Int. Conf. Field-Programmable Logic Appl. (FPL), Dresden, Germany, 2021, pp. 69–75.

S. Haag and R. Anderl, "Digital twin—proof of concept", Manufacturing Letters, vol. 15, pp. 64–66, Jan. 2018.

P. Hitzler, A. Eberhart, M. Ebrahimi, M. K. Sarker, and L. Zhou, "Neuro-symbolic approaches in artificial intelligence", National Science Review, vol. 9, no. 6, Jun. 2022.

A. Letichevsky, "Algebra of behavior transformations and its applications", in Structural Theory of Automata, Semigroups, and Universal Algebra, V. B. Kudryavtsev and I. G. Rosenberg, Eds. Dordrecht, The Netherlands: Springer, 2005, pp. 241–272.

O. Letychevskyi and V. Peschanenko, "Applying algebraic virtual machine to cybersecurity tasks", in Proc. 2022 IEEE 9th Int. Conf. Sci. Electron., Technol. Inf. Telecommun. (SETIT), Hammamet, Tunisia, 2022, pp. 161–169.

Y. Mirsky, "Kitsune network attack dataset", Kaggle.


Refbacks

  • There are currently no refbacks.