DOI
10.34229/KCA2522-9664.25.2.1
UDC 004.05
|
|
2 V.M. Glushkov Institute of Cybernetics, National Academy of Sciences of Ukraine, Kyiv, Ukraine
 bogdanscloud@gmail.com
|
THE PROBLEM OF ACCURACY IN SYSTEMS FOR CYBERATTACK
RESISTANCE AND THE VERIFICATION OF NEURAL NETWORKS
ON THE EXAMPLE OF BOTNET DETECTING PROBLEM
Abstract. The paper considers the problem of the accuracy of detection of intrusions in software systems based on deep learning neural networks. An example of a system for detecting botnets, malicious software, which are the source of potential attacks, including Denial of Service (DDoS), is presented. The system is created as a classification model that detects the behavior of botnets on infected resources. A number of experiments were conducted on the open data set of the Canadian Institute for Cybersecurity. To increase the accuracy of the classification, the method of augmentation of the data set using the method of generating examples of adversarial attacks was used. A method for verification of the reliability of a neural network using automatic proof of the robustness property of the model based on SMT solvers is presented. To increase the accuracy of attack detection, a neurosymbolic approach that combines algebraic methods with classification models is also considered.
Keywords: cyber security, botnet, algebraic modelling, deep learning neural network, adversarial attacks, verification.
full text
REFERENCES
- 1 Kabay M.E. Kraken the botnet: The ethics of counter-hacking. NetworkWorld; Southborough, 2009. https://www.networkworld.com/article/ .
- 2. Open-appsec by Checkpoint. https://www.openappsec.io/ .
.
- 3. Imperva Advanced Bot Protection. https://www.imperva.com/products/ .
- 4. NETACEA. AI-driven bot protection. Block bots effortlessly. https://netacea.com/ .
- 5. AWS WAF bot control. https://docs.aws.amazon.com/waf/ .
- 6. F5 distributed cloud bot defense. https://www.f5.com/products/ .
- 7. Liu Y., Mao S., Mei X., Yang T., Zhao X. Sensitivity of adversarial perturbation in fast gradient sign method. Proc. 2019 IEEE Symposium Series on Computational Intelligence (SSCI) (06-09 December 2019, Xiamen, China). Xiamen, 2019. P. 433–436. https://doi.org/10.1109/SSCI44817.2019.9002856.
- 8. Casadio M., Komendantskaya E., Daggitt M.L., Kokke W., Katz G., Amir G., Refaeli I. Neural network robustness as a verification property: A principled case study. arXiv:2104.01396v2 [cs.LG] 13 Jul 2022. https://doi.org/10.48550/ .
- 9. Balan R., Singh M., Zou D. Lipschitz properties for deep convolutional networks. arXiv:1701.05217v1 [cs.LG] 18 Jan 2017. https://doi.org/10.4855/arXiv.1701.05217.
- 10. Letichevsky A. Algebra of behavior transformations and its applications. Proc. NATO Advanced Study Institute on Structural Theory of Automata, Semigroups and Universal Algebra (7–18 July 2003, Montreal, Quebec, Canada). Montreal, 2003. NATO science series II: Mathematics, physics and chemistry. Dordrecht: Springer, 2005. Vol. 207. P. 241–272. https://doi.org/10.1007/ .
- 11. Letychevskyi O., Peschanenko V. Applying algebraic virtual machine to cybersecurity tasks. Proc. 2022 IEEE 9th International Conference on Sciences of Electronics, Technologies of Information and Telecommunications (SETIT) (28–30 May 2022, Hammamet, Tunisia). Hammamet, 2022. P. 161–169. https://doi.org/10.1109/SETIT54465.2022.9875895 .
