Abstract
In the paper, the problem of the possibility of recovering the unknown structure of artificial neural networks (ANNs) using the theory of graphs is investigated. The key ANN concepts, their typical architectures, and differences are considered. The application of the theory of the graph tool for solving the problem of detecting an ANN structure is substantiated, and examples of comparing different ANN architectures and graph types are presented. It is proposed to use the methods of the spectral graph theory and the graph signal processing as tools for analyzing the ANN structure.
Similar content being viewed by others
REFERENCES
Vasiliev, P.D., Zegzhda, and Kuvshinov, V.I., Modern problems of cybersecurity, Nonlinear Phenom. Complex Syst. (Dordrecht, Neth.), 2014, vol. 17, no. 3, pp. 210–214.
Poltavtseva, M.A. and Kalinin, M.O., Conceptual data modeling using aggregates to ensure large-scale distributed data management systems security, Stud. Comput. Intell., 2020, vol. 868, pp. 41–47.
Platonov, V.V. and Semenov, P.O., Using data-mining methods to detect network attacks, Autom. Control Comput. Sci., 2015, vol. 49, no. 8, pp. 766–769.
Malyshev, E.V., Moskvin, D.A., and Zegzhda, D.P., Application of an artificial neural network for detection of attacks in VANETs, Autom. Control Comput. Sci., vol. 53, no. 8, pp. 889–894. https://doi.org/10.3103/S0146411619080194
Krundyshev, V., Kalinin, M., and Zegzhda, P., Artificial swarm algorithm for VANET protection against routing attacks, 2018 IEEE Industrial Cyber-Physical Systems, 2018, pp. 795–800, https://doi.org/10.1109/ICPHYS.2018.8390808
Zegzhda, P.D., Malyshev, E.V., and Pavlenko, E.Y., The use of an artificial neural network to detect automatically managed accounts in social networks, Autom. Control Comput. Sci., 2017, vol. 51, no. 8, pp. 874–880.
Markov, Y.A. and Kalinin, M.O., Intellectual intrusion detection with sequences alignment methods, Lect. Notes Comput. Sci., 2010, vol. 6258, pp. 217–228.
Belenko, V., Krundyshev, V., and Kalinin, M., Intrusion detection for Internet of Things applying metagenome fast analysis, 3rd World Conference on Smart Trends in Systems, Security and Sustainability, WorldS4, 2019, pp. 129–135. https://doi.org/10.1109/WorldS4.2019.8904022
Belenko, V., Chernenko, V., Kalinin, M., and Krundyshev, V., Evaluation of GAN applicability for intrusion detection in self-organizing networks of cyber physical systems, 2018 International Russian Automation Conference (RusAutoCon), 2018. https://doi.org/10.1109/RUSAUTOCON.2018.8501783
Krundyshev, V. and Kalinin, M., Hybrid neural network frame work for detection of cyber attacks at smart infrastructures, ACM International Conference Proceeding Series, 2019. https://doi.org/10.1145/3357613.3357623
Kalinin, M., Demidov, R., and Zegzhda, P., Hybrid neural network model for protection of dynamic cyber infrastructure, Nonlinear Phenom. Complex Syst. (Dordrecht, Neth.), 2019, vol. 22, no. 4, pp. 375–382.
Zegzhda, P.D., Zegzhda, D.P., and Nikolskiy, A.V., Using graph theory for cloud system security modeling, Lect. Notes Comput. Sci., 2012, vol. 7531, pp. 309–318. https://doi.org/10.1007/978-3-642-33704-8-26
Ivanov, D.V. and Moskvin, D.A., Application of fractal methods to ensure the cyber-resilience of self-organizing networks, Nonlinear Phenom. Complex Syst. (Dordrecht, Neth.), 2019, vol. 22, no. 4, pp. 336–341.
Pavlenko, E.Y., Yarmak, A.V., and Moskvin, D.A., Application of clustering methods for analyzing the security of android applications, Autom. Control Comput. Sci., 2017, vol. 51, no. 8, pp. 867–873.
Zegzhda, P., Zegzhda, D., Pavlenko, E., and Dremov, A., Detecting Android application malicious behaviors based on the analysis of control flows and data flows, ACM International Conference Proceeding Series, 2017, pp. 280–286. https://doi.org/10.1145/3136825.3140583
Zegzhda, P., Zegzhda, D., Pavlenko, E., and Ignatev, G., Applying deep learning techniques for Android malware detection, ACM International Conference Proceeding Series, 2018. https://doi.org/10.1145/3264437.3264476
Yuan, X., et al., Adversarial examples: Attacks and defenses for deep learning, IEEE Trans. Neural Networks Learn. Syst., 2019, vol. 30, no. 9, pp. 2805–2824.
Zhang, J. and Li, C., Adversarial examples: Opportunities and challenges, IEEE Trans. Neural Networks Learn. Syst., 2019.
Oh, S.J., Schiele, B., and Fritz, M., Towards reverse-engineering black-box neural networks, in Explainable AI: Interpreting, Explaining and Visualizing Deep Learning, Cham: Springer, 2019, pp. 121–144.
Nikolenko, S., Lecture 11. Neural Networks. https://logic.pdmi.ras.ru/~sergey/oldsite/teaching/asr/notes-11-neural.pdf.
The Azimov Institute. The Neural Network Zoo. https://www.asimovinstitute.org/neural-network-zoo.
Chung, F.R.K., Lectures on spectral graph theory, CBMS Lect. (Fresno), 1996, vol. 6, pp. 17–21.
Gripon, V., Ortega, A., and Girault, B., An inside look at deep neural networks using graph signal processing, 2018 Information Theory and Applications Workshop (ITA), 2018, pp. 1–9.
Anirudh, R., et al., Margin: Uncovering Deep Neural Networks Using Graph Signal Analysis, 2017. arXiv:1711.05407.
Funding
The study was carried out within the State Assignment for Basic Research, project no. 0784-2020-0026, and was supported by the Collateral Agreement to the Agreement on the Grant from the Federal Budget for Financial Support of the State Assignment for the Provision of Public Services (execution of works) no. 075-03-2020-158/2 dated March 17, 2020 (internal no. 075-GZ/Shch4575/784/2).
Author information
Authors and Affiliations
Corresponding authors
Ethics declarations
The authors declare that they have no conflicts of interest.
Additional information
Translated by N. Semenova
About this article
Cite this article
Lavrova, D.S., Shtyrkina, A.A. The Analysis of Artificial Neural Network Structure Recovery Possibilities Based on the Theory of Graphs. Aut. Control Comp. Sci. 54, 977–982 (2020). https://doi.org/10.3103/S0146411620080222
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.3103/S0146411620080222