Optimization of the evolutionary matching method based on the probability of an erroneous decision

Roman Mirakhmedov, Zinaida Potapova

Abstract


The problem of obtaining demonstrably correct solutions to problems posed to experts is discussed and a methodology for solving local problems with a predetermined low error probability by a team of experts is presented. To build a theory of systems of evolutionary coordination, the following were used: Turchin’s theory of metasystem transitions; Condorcet's jury theorem; the Rasch model, which takes into account the connection between the probability of a correct decision and the actor’s preparedness and the difficulty of the task; ternary logic, which allows experts to accept the answer “I don’t know” in difficult cases; multi-stage work of actors acting as solution generators and expert evaluators of other people's solutions with coordination of their work by a genetic algorithm. A theoretical study of variations of the method of evolutionary coordination of solutions was carried out. The paper presents and substantiates mathematical models of variations of the method of evolutionary coordination of decisions, and proves a theorem about the tendency of the probability of an erroneous decision to zero when a number of conditions are met. The results of computer modeling of the decision-making process using the Monte Carlo method are presented. The results of computer modeling coincide with the results of the theoretical model within the limits of statistical errors. It is concluded that all the declared properties of systems of evolutionary coordination of decisions, if possible, to give either the correct answer with a given value of the probability of an erroneous decision, or the answer “I don’t know,” can be fulfilled by choosing a type of model, the number of actors and the rank of the actor in accordance with the local task.

Full Text:

PDF (Russian)

References


Gorban A. N.: Errors in data-based intelligence. Collection of articles based on the materials of the International Conference “Intelligent Systems in Science and Technology. artificial intelligence in solving pressing social and economic problems of the 21st century.” – Perm: –2020, pp. 11–13.

Fields, Chris, James F. Glazebrook, and Michael Levin.. Principled Limitations on Self-Representation for Generic Physical Systems // Entropy. –2024. Vol. 3. P. 1–16.

Keeling, S.: Confabulation and rational obligations for self-knowledge // Philos. Psychol. 2018, 31, pp.1215–1238. https://doi.org/10.1080/09515089.2018.1484086

G. Vilone, L. Longo. “Notions of explainability and evaluation approaches for explainable artificial intelligence”, Information Fusion, vol. 76, pp. 89-106, Dec. 2021. doi:10.1016/j.inffus.2021.05.009.

V. Protasov, Z. Potapova , G. Akhobadze. “How to reduce the probability of erroneous decisions in the systems of collective intelligence”, IOP Conference Series: Materials Science and Engineering, 2020, v.927(1), 012069.

https://iopscience.iop.org/article/10.1088/1757-899X/927/1/012069/pdf

Protasov V.I. Collective intelligence systems. Theory and practice. – M: University book. – 2024. – 230 p. (in Russian).

L.V. Markaryan. Models and algorithms of collective intelligence systems. – M: Ed. MISiS. – 2020. – 104 p. (in Russian).

Alexander Gorban , Bogdan Grechuk, Ivan Tyukin.: Augmented Artificial Intelligence: a Conceptual Framework. 18.03.2018, arXiv preprint arXiv: https://arxiv.org/pdf/1802.02172v3.pdf (дата обращения 21.03.2024).

V.F. Turchin. Phenomenon of science. Cybernetic approach to evolution. –M.: Sinteg, 1993, 456 p. (in Russian).

V.I. Protasov. Construction of metasystem transitions. –M.: ICPT Publishing House, 2009, 186 p. (in Russian).

Сondorcet, marquis Marie-Jean-Antoine-Nicolas de Caritat. “Essai sur l’application de l’analyse à la probabilité des décisions rendues à la pluralité des voix” . –Paris: Imprimerie Royale, 1785.

G. Rasch. Probabilistic Models for Some Intelligence and Attainment Tests. –Chicago: University of Chicago Press, 1981,199 p.

Protasov V.I. Method of evolutionary coordination of solutions. Computer and mathematical models. Mining information and analytical bulletin. – Moscow: Mining Book Publishing House. – 2011. – No. 1. – P. 360-379 (in Russian).

V.I. Protasov, Z.E. Potapova. “Drawing up an identikit using the method of intellectual consultation.” World of Measurements, M.: Standards and Quality, 2009, No. 3, p. 10-21. (in Russian).

V.I. Protasov, P.D. Rabinovich. Application of new network information technology of an intelligent council to compile an identikit. Proceedings of the Institute for Systems Analysis. - Moscow. – 2010. – vol. 1. – P. 279-286. (in Russian).

V.I. Protasov. “Application of the network method of evolutionary coordination of decisions in project management.” Project and program management. M.: Grebennikov Publishing House, 2011, t. 1(25), p. 22–35. (in Russian).

V.I. Protasov. “Setting the parameters of the method of evolutionary coordination of solutions” News of the Kabardino-Balkarian Scientific Center of the Russian Academy of Sciences, 2012, No. 2(46), pp. 29–37 (in Russian).

M. Mirakhmedov, Z. Potapova, V. Protasov. “MESING – a new method of organizing the joint work of neural networks and its metrology”. Journal of Physics: Conference Series , 2021, v. 1727, 012004. https://iopscience.iop.org/article/10.1088/1742-6596/1727/1/012004/pdf (дата обращения 21.03.2024).

V.I. Protasov, Z.E. Potapova. “Methodology for radically reducing the likelihood of making erroneous decisions in collective intelligence systems.” International scientific journal "Modern information technologies and IT education". 2019, volume 15, no. 3, p. 588 – 601. (in Russian).

V.I. Protasov, Z.E. Potapova. “Metrology of systems of evolutionary coordination of decisions and rationing of intellectual work.” International scientific journal "Modern information technologies and IT education". 2019, volume 15, no. 4, h. 1047-1055. (in Russian)


Refbacks

  • There are currently no refbacks.


Abava  Кибербезопасность MoNeTec 2024

ISSN: 2307-8162