Two-way attack on synchronized artificial neural networks forming a common secret
https://doi.org/10.21122/2309-4923-2025-2-55-59
Abstract
The vulnerability of the technology for generating a common cryptographic key using synchronized artificial neural networks (ANN) is considered in a relatively new type of attack called two-way. The essence of the attack is that a cryptanalyst, listening to an open communication channel, creates two identical artificial neural networks, one of which he synchronizes with the network of one of the legal subscribers, and the second with the network of another legal subscriber. By comparing the vectors of weight coefficients of attacking networks, the cryptanalyst is able to determine the moment of full synchronization of the networks of legal subscribers and the value of the generated secret number. Next, we examine the possibilities of a two-way attack using various models of secret number generation.
About the Authors
U. F. HolikauBelarus
Uladzimir F. Holikau, doctor of Technical Sciences, Professor. Area of scientific interests: infor- mation security, cryptography.
Minsk, Republic of Belarus.
M. L. Radziukevich
Scientific Production Republican Unitary Enterprise "Research Insti- tute for Technical Protection of Information"
Belarus
Belarus
Maryna L. Radziukevich, Ph.D. of Engineering Sciences, head of the testing laboratory for information security requirements.
Minsk, Republic of Belarus
D. V. Shuliak
Scientific Production Republican Unitary Enterprise "Research Institute for Technical Protection of Information"
Belarus
Belarus
Dmitry V. Shuliak, Scientific. Master of Engineering Sci- ence, leading engineer of the testing laboratory for information security requirements .
Minsk, Republic of Belarus
References
1. Kanter, I. The Theory of Neural Networks and Cryptography / Ido Kanter, Wolfgang Kinzel // The Phisics of Communication. – 2003. – P. 631–642. – DOI: 10.1142/9789812704634_0044
2. Kinzel, W. Neural Cryptography / W. Kinzel, I. Kanter // ICONIP'02 : proc. of the 9th Intern. Conf. on Neural Information Processing, Nov. 18–22, 2002, Orchid Country Club, Singapore. – Singapore, 2002. – Vol. 3. – P. 1351–1354. – DOI: 10.1109/ICONIP.2002.1202841
3. Ruttor, A. Dynamics of neural cryptography / A. Ruttor, I. Kanter, W. Kinzel // Physical Review E – Statistical, Nonlinear, and Soft Matter Physics. – 2007. – Vol. 75, iss. 5. – P. 056104.
4. Urbanovich, P.P. Usage of hypercomplex numbers in a cryptographic key agreement protocol based on neural networks / P.P. Urbanovich, N.P. Shutko // Journal of the Belarusian State University. Mathematics and Informatics. – 2024. – Vol. 2. – P. 81–92.
5. Golikov, V.F. Attack on synchronized artificial neural networks, forming a common secret by deferred search /V.F. Golikov, A.Y. Ksenevich // Doklady BGUIR. – 2017. – Vol. 8. – P. 48-53. (In Russ.)
6. Radziukevich, M.L. Enhancing the secrecy of a cryptographic key generated using synchronized artificial neural networks / M.L. Radziukevich, V.F. Golikov // Informatics. – 2020. – Vol. 17, № 1. – P. 102-108. – DOI: 10.37661/1816-0301-2020-17-1-102-108
7. Radziukevich, M.L. Combined formаtion of a cryptographic key using synchronized artificial neural networks / M.L. Radziukevich, V.F. Golikov // Doklady BGUIR. – 2021. – Vol. 19, № 1. –P. 79-87. – DOI: 10.35596/1729-7648-2021-19-1-79-87
8. Radziukevich, M.L. A combined method of formation of a cryptographic key with secret modification of the results of synchronization of artificial neural networks. System analysis and applied information science. – 2021. – Vol. 3. – P.51-58. (In Russ.). – DOI: 10.21122/2309-4923-2021-3-51-58
Review
For citations:
Holikau U.F., Radziukevich M.L., Shuliak D.V. Two-way attack on synchronized artificial neural networks forming a common secret. «System analysis and applied information science». 2025;(2):55-59. (In Russ.) https://doi.org/10.21122/2309-4923-2025-2-55-59
Views:
9
Email this article 