CONCEPTUAL FUNDAMENTALS OF DEPENDABLE ARTIFICIAL INTELLIGENCE SYSTEMS
DOI:
https://doi.org/10.15407/dopovidi2025.02.011Keywords:
dependable computing, trustworthy artificial intelligence, AI system quality model, von Neumann’s paradigm, principle of diversityAbstract
The concept of dependent artificial intelligence (AI) systems based on the development of the von Neumann paradigm (VNP) is proposed. It is presented by a set-theoretic description that takes into account different qualitative characteristics of AI and AI systems (AIS). The stages and formulations of the evolution of GNPs from simple relay units to complex digital infrastructures and AIS are analyzed. One of the stages of GNP development is related to the fundamental work on the concepts and taxonomy of reliable and secure computing (A. Avizienis et. al., 2004). The AIS Quality Model (QM) is described as an ordered hierarchy of attributes (characteristics) of trustworthiness, explainability, ethicality, legality, responsibility and their specific sub-characteristics, which allows to determine the possibilities of applying GNP to ensure the required values of the characteristics. VNP is formulated for AIS in various representations such as “ trustworthy AIS form untrustworthy components”. AIS QM consists of the AI quality model and the QM of the system’s hardware-software platform. Application examples of AIS QM are analyzed. A model for matching and transforming input data into AIS output data is proposed, taking into account the decomposition of a universal data set into subsets used for training and possible anomalies in certain quality characteristics, as well as different types of failures and cyberattacks on AIS. It is proposed to use the principle of diversity in the implementation of VNP to ensure reliability and other characteristics of AI and to create dependent AIS. Models of reliable multiversion AIS are described and methods of reliability improvement are considered. These methods are based on various online testing schemes and application of versioning and structural redundancy.
Downloads
References
Ding, W., Abdel-Basset, M., Hawash, H. & Ali, A. M. (2022). Explainability of AI methods, applications and challenges: A comprehensive survey. Information Science, 615 (C), рр. 238-292. https://doi.org/10.1016/j. ins.2022.10.013
Kharchenko, V., Fesenko, H. & Illiashenko, O. (2022). Basic model of non-functional characteristics for assessment of artificial intelligence quality. Radioelectron. Comput. Syst., No. 2, рр. 1-14. https://doi. org/10.32620/reks.2022.2.11
Kharchenko, V., Fesenko, H. & Illiashenko O. (2022). Quality Models for Artificial Intelligence Systems: Characteristic-Based Approach, Development and Application, 22(13), рр. 1-32. https://doi.org/10.3390/ s22134865
Williams, R. & Yampolskiy, R. (2021). Understanding and Avoiding AI Failures: A Practical Guide. Philosophies, 53, 6, рр. 1-25. https://doi.org/10.3390/philosophies6030053
Steimers, A. & Schneider, M. (2022). Sources of Risk of AI Systems. Int. J. Environ. Res. Public Health, 19, 3641, рр. 1-26. https://doi.org/10.3390/ijerph19063641
Ziwei, Xu, Sanjay, Jain & Mohan Kankanhalli. (2022). Hallucination is Inevitable: An Innate Limitation of Large Language, 22, No. 1, рр.1-26. https://doi.org/10.48550/arXiv.2401.11817
Avizienis, A., Laprie, J.-C., Randell, B. & Landwehr, C. (2004). Basic concepts and taxonomy of dependable and secure computing. IEEE Trans. Dependable Secur. Comput., 1, рр. 11-33. https://doi.org/10.1109/TDSC.2004.2
Kharchenko, V. S. (2009). Dependable systems and multi-version computing: evolution issues. Radioelecntronni ta compjuterni systemy, No. 7, рр. 46-59 (in Ukrainian).
Kharchenko, V. S. (2007). Dependability of computer systems: boarder of universality in context of information- technical states. Radioelecntronni ta compjuterni systemy, No. 8, рр. 7-14 (in Ukrainian).
Nikolas, Guelfi. (2011). A Formal Framework for Dependability and Resilience from a Software Engineering Perspective. Cent. Eur. J. Comp. Sci., 1(3), рр. 294-328. https://doi.org/:10.2478/s13537-011-0025-x
Moskalenko, V., Kharchenko, V., Moskalenko, A. & Kuzikov, B. (2023). Resilience and Resilient Systems of Artificial Intelligence: Taxonomy, Models and Methods. Algorithms, 16, 165, рр. 1-34. https://doi.org/10.3390/ a16030165
Neumann, J. von. (1956). Probabilistic Logics and the Synthesis of Reliable Organisms from Unreliable Components. Ed. by C. E. Shannon, J. McCarthy, Princeton University Press, рр. 43-98. https://doi. org/10.1515/9781400882618-003
Kharchenko, V. & Gorbenko, A. (2013). Evolution of von Neumann’s paradigm: Dependable and green computing, East-West Design & Test Symposium, рр. 1-6. https://doi.org/10.1109/EWDTS.2013.6673090
Kharchenko, V. & Odarushchenko, O. (2024). Trustworthy AI Systems from Untrustworthy Components: Development von Neumann’s Paradigm using Principle of Diversity. Proceedings 4th International Workshop of IT-professionals on Artificial Intelligence (ProfIT AI 2024), September 25–27, Cambridge, MA, USA, рр. 392-404.
Veprytska, O. & Kharchenko, V. (2022). AI powered attacks against AI powered protection: classification, scenarios and risk analysis, 12th International Conference on Dependable Systems, Services and Technologies (DESSERT), IEEE, рр. 1-7. https://doi.org/10.1109/dessert58054.2022.10018770
Illiashenko, O., Kharchenko, V., Babeshko, I., Fesenko, H. & Di, F. (2023). Giandomenico, Security-Informed Safety Analysis of Autonomous Transport Systems Considering AI-Powered Cyberattacks and Protection. Entropy, 25, No. 8, pp.1-35. https://doi.org/:10.3390/e25081123
Veprytska, O. & Kharchenko, V. (2024). Analysis of AI powered attacks and protection of UAV assets: quality model-based assessing cybersecurity of mobile system for demining Proceeding of 5th International Workshop on Intelligent Information Technologies and Systems of Information Security, March 28, Khmelnytskyi, Ukraine. P. 356-374.
Moskalenko, V. & Kharchenko, V. (2024). Resilience-aware MLOps for AI-based medical diagnostic system Front. Public Health, 27 March 2024 Sec. Disaster and Emergency Medicine, Vol. 12, рр. 1-12. https://doi. org/10.3389/fpubh.2024.1342937
Neretin, O., Kharchenko, V. & Fesenko, H. (2023). Multi-source Analysis of AI Vulnerabilities: Methodology and Algorithms of Data Collection, IEEE 12th International Conference on Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications (IDAACS), Dortmund, Germany, рр. 972-977. https://doi.org/10.1109/IDAACS58523.2023.10348671
Babeshko, I., Illiashenko, O., Kharchenko, V. & Leontiev, K. (2022). Towards Trustworthy Safety Assessment by Providing Expert and Tool-Based XMECA Techniques. Mathematics. 10, рр. 1-29. https://doi.org/10.3390/ math10132297
Siora, А. А., Krasnobaev, V. А. & Kharchenko, V. S. (2009). Fault-tolerant systems with version-information redundancy. Ed. V. S. Kharchenko. Kharkiv: National aerospace university «KhAI», 321 p. (in Russian).
Kharchenko, V. S., Odarushchenko, O. M., Fesenko, H. V. & Odarushchenko, O. B. Method of intelligent system reservation: patent on know-how model № 156654, request № u202304653; submit. 03.10.2023; publish. 24.07.2024, bul. № 30 (in Ukrainian).
Fedorenko, G., Fesenko, H., Kharchenko, V., Kliushnikov, I. & Tolkunov, I. (2023). Robotic-biological systems for detection and identification of explosive ordnance: Concept, general structure, and models. Radioelectron. Comput. Syst., No. 2, рр. 143-159.
Fedorenko, G.., Kliushnikov, I., Kharchenko, V. et al. Method of search and recognition of explosive ordnance: patent on know-how model № 1542266, request № u202300129; submit. 03.01.2023; publish. 25.10.2023, bul.
№ 43 (in Ukrainian).
Fesenko, H., Illiashenko, O., Kharchenko, V., Kliushnikov, I., Morozova, O., Sachenko, A. & Skorobohatko, S. (2023). Flying Sensor and Edge Network-Based Advanced Air Mobility Systems: Reliability Analysis and Applications for Urban Monitoring. Drones, No. 7, рр. 409. https://doi.org/10.3390/drones7070409
Ivanchenko, O., Kharchenko, V., Smyrynska, N. & Veprytska, О. (2024). Cybersecurity of unmanned surface vessels: IMECA based assessment and protection against ai powered attacks. Annual of Navigation. https://doi. org/10.36163/aon-2024-0004
Mishchuk, V., Fesenko, H. & Kharchenko, V. (2024). Deep learning models for detection of explosive ordnance using autonomous robotic systems: trade-off between accuracy and real-time processing speed. Radioelectronic and Computer Systems, No. 4, рр. 99-111. https://doi.org/10.32620/reks.2024.4.09
Ponochovnyi, Y. & Kharchenko, V. (2020). Dependability Assurance Methodology of Information and Control Systems using multipurpose service strategies. Radioelectron. Comput. Syst., 3 ,рр. 43-58. https://doi. org/10.32620/reks.2020.3.05
Moskalenko, V., Kharchenko, V. & Semenov, S. (2024). Model and Method for Providing Resilience to Resource- Constrained AI-System. Sensors, 24, 5951. https://doi.org/10.3390/s24185951
Golembovska Olena & Kharchenko Vyacheslav. (2025). Technologies of Interactive Art Encyclopedia of Libraries, Librarianship, and Information Science. Elsevier, рp. 506-527. https://doi.org/10.1016/b978-0-323- 95689-5.00152-8
Barmak, O., Krak, I., Yakovlev, S., Radiuk, P. & Kuznetsov, V. (2024). Toward explainable deep learning in healthcare through transition matrix and user-friendly features. Frontiers in Artificial Intelligence, 7, 1482141. https://doi.org/10.3389/frai.2024.1482141
Vakulenko, D. V., Palagin, O. V., Sergienko, I. V. et al. (2024). Algorithmization and Optimization Models of Patient-Centric Rehabilitation Programs. Cybernetics and Systems Analysis, 60, рр. 736-752. https://doi. org/10.1007/s10559-024-00711-5
Veprytska, O. & Kharchenko, V. (2022). Analysis of requirements and quality model-oriented assessment of explainable AI as a service. Electronic Modeling, 44, No. 5, pp. 36-50. https://doi.org/10.15407/emodel.44.05.036
Palagin, O., Kaverinskiy, V., Malakhov, K. & Petrenko, M. (2024). Fundamentals of the Integrated Use of Neural Network and Ontolinguistic Paradigms: A Comprehensive Approach. Cybern. Syst. Anal., 60, No. 1, pp. 111- 123. https://doi.org/10.1007/s10559-024-00652-z
Sergienko, I. V., Molchanov, I. N. & Khimich, A. N. (2010). Intelligent technologies of high-performance computing. Cyber. Syst. Anal., 46, No. 5, рр. 833-844. https://doi.org/10.1007/s10559-010-9265-3
Gordieiev, O., Rainer, A., Kharchenko, V., Pishchukhina, O. & Gordieieva, D. (2024). A Unified Approach to the Development of Technology-Based Software Quality Models on the Example of Blockchain Systems. IEEE Access, 12, рр. 118875-118889. https://doi.org/10.1109/ACCESS.2024.3448271
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2025 Reports of the National Academy of Sciences of Ukraine
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
