Adaptive symmetric cryptosystem with dynamic key agreement based on deep neural networks
DOI:
https://doi.org/10.33216/1998-7927-2025-296-10-5-12Keywords:
adaptive cryptosystem, symmetric encryption, dynamic key agreement, deep neural networks, neural synchronization, cryptographic robustnessAbstract
The article is devoted to the development of an adaptive symmetric cryptosystem. Key formation is performed without traditional key exchange protocols. The method is based on the synchronization of deep neural networks. The process is implemented over an open communication channel. The purpose of the study is to substantiate a new architecture for dynamic key management in the space of neural networks. The mechanism relies on neural synchronization. Cryptographic robustness and computational efficiency of the model are evaluated. The methodology integrates cryptographic analysis and neural network theory. A formalized description of algorithmic constructions is applied. An analytical assessment of computational complexity is carried out. A comparison with classical symmetric encryption schemes is performed. The result is an architecture with decentralized key agreement logic. The key is generated as a function of the internal state of the network. The algorithm covers initialization, synchronization, and key generation. The system provides cyclic updating of key material. Encryption and decryption procedures are implemented. Resistance to passive attacks is analyzed. Robustness against active forms of interference is investigated. The system counteracts interception and injection attacks. The mechanism protects against man-in-the-middle attacks. Model adaptivity reduces correlation between successive keys. Reconstruction of internal states is significantly complicated. Computational complexity remains acceptable for practical application. The neural component does not reduce system performance during long communication sessions. Practical value lies in application within distributed networks. The system operates without centralized trust infrastructure. The results are suitable for embedded computing systems. The approach is applicable to mobile and sensor networks. The model reduces dependence on asymmetric cryptography. Peak computational loads are eliminated. Future work is associated with experimental verification on hardware platforms. An in-depth analysis of resistance to combined attack vectors is planned. A separate scientific interest concerns optimization of architectural solutions.
References
1. Abrar A., Dasgupta S., Rahman M., Alsharif A. AI-driven post-quantum cryptography for cyber-resilient V2X communication in transportation cyber-physical systems. arXiv. 2025. https://doi.org/10.48550/arXiv.2510.08496
2. Alanazi M. J., Alhoweiti R. A., Alhwaity G. A., Alharbi A. R. An adaptive hybrid cryptographic framework for resource-constrained IoT devices. Electronics. 2025. Vol. 14. No. 23. Article 4666. https://doi.org/10.3390/electronics14234666
3. Alzaidy S., Binsalleeh H. Adversarial attacks with defense mechanisms on convolutional neural networks and recurrent neural networks for malware classification. Applied Sciences. 2024. Vol. 14. No. 4. Article 1673. https://doi.org/10.3390/app14041673
4. Bhat R., Nanjundegowda R. Exploring generative adversarial networks for secure data encryption and future directions in communication systems. Proceedings of the International Conference on Futuristic Technology. 2025. https://doi.org/10.5220/0013591300004664
5. Geng Y. Identification of cryptosystem based on deep neural network. Highlights in Science, Engineering and Technology. 2025. Vol. 142. p. 391–399. https://doi.org/10.54097/zfm1nx90
6. Hanafi B., Bokhari M. U., Wani M. A., Shakil K. A., Ali G. Dynamic adversarial neural cryptography for ensuring privacy in smart contracts. PeerJ Computer Science. 2025. Vol. 11. Article e3286. https://doi.org/10.7717/peerj-cs.3286
7. Hao J., Jin M., Li Y., Yang Y. Neural network-based symmetric encryption algorithm with encrypted traffic protocol identification. PeerJ Computer Science. 2025. Vol. 11. Article e2750. https://doi.org/10.7717/peerj-cs.2750
8. He Z., Sayadi H. AI in chaos: Adaptive and secure communication via deep reinforcement learning and moving target defense. IEEE Access. 2025. Vol. 13. p. 199971–200000. https://doi.org/10.1109/ACCESS.2025.3635665
9. Jain K. Exploring cryptographic key management schemes for enhanced security in WSNs. Journal of Information and Applied Science. 2025. Vol. 20. No. 1. p. 18–37. https://doi.org/10.2478/ias-2025-0002
10. Jung I. S., Song Y. R., Jilcha L. A., Kim D. H., Im S. Y., Shim S. W., Kim Y. H., Kwak J. Enhanced encrypted traffic analysis leveraging graph neural networks and optimized feature dimensionality reduction. Symmetry. 2024. Vol. 16. No. 6. Article 733. https://doi.org/10.3390/sym16060733
11. Koshiba T., Zolfaghari B., Bibak K. A tradeoff paradigm shift in cryptographically-secure pseudorandom number generation based on discrete logarithm. Journal of Information Security and Applications. 2023. Vol. 73. Article 103430. https://doi.org/10.1016/j.jisa.2023.103430
12. Kumar P. R., Goel S. A secure and efficient encryption system based on adaptive and machine learning for securing data in fog computing. Scientific Reports. 2025. Vol. 15. Article 11654. https://doi.org/10.1038/s41598-025-92245-9
13. Makedon V. V., Kholod O. H., Yarmolenko L. I. The model for assessing the competitiveness of high tech enterprises on the basis of the formation of key competences. Academic Review. 2023. Vol. 59. No. 2. p. 75–89. https://doi.org/10.32342/20745354-2023-2-59-5
14. Makedon V., Myachin V., Aloshyna T., Cherniavska I., Karavan N. Improving the readiness of enterprises to develop sustainable innovation strategies through fuzzy logic models. Economic Studies (Ikonomicheski Izsledvania). 2025. Vol. 34. No. 5. p. 165–179. URL: https://archive.econ-studies.iki.bas.bg/2025/2025_05/2025_05_09.pdf
15. Mohsin Z. R. AI-powered encryption revolutionizing cybersecurity with adaptive cryptographic algorithms. Turkish Journal of Computer and Mathematics Education. 2025. Vol. 16. No. 1. p. 44–62. https://doi.org/10.61841/turcomat.v16i1.14976
16. Reddy P. S., Reddy T. S., Khaled M. K. Enhanced steganography using dynamic compression and encryption algorithms. International Journal of Engineering Innovations and Management Strategies. 2025. Vol. 1. No. 3. p. 1–13.
17. Saha A., Pathak C., Saha S. A study of machine learning techniques in cryptography for cybersecurity. American Journal of Electronics & Communication. 2021. Vol. 1. No. 4. p. 22–26.
18. Singh M., Baranwal N., Singh K. N., Singh A. K. Using GAN-based encryption to secure digital images with reconstruction through customized super resolution network. IEEE Transactions on Consumer Electronics. 2023. Vol. 70. No. 1. p. 3977–3984. https://doi.org/10.1109/TCE.2023. 3276732
19. Zarei M., Dindarlou M. H. F., Taghizadeh M., et al. Lightweight image encryption for wireless sensor networks using optimized elliptic curve and fuzzy logic. Scientific Reports. 2025. https://doi.org/10.1038/s41598-025-32877-z
