UDC 004.8
DOI: 10.36871/2618-9976.2023.10-2.003
Authors
Svetlana V. Prokopchina,
Doctor of Technical Sciences, Professor, Financial University under the Government of the Russian
Federation, Moscow, Russia
Abstract
The article is devoted to a new direction in the creation of information and measurement systems – measuring artificial intelligence. The paper gives the basic definitions and attributes of measuring artificial intelligence. The field of measurement tasks in which the application of methods and means of intellectualization of measurement processes is necessary is defined. The main types of intelligent measurements are given, their characteristics and specific properties are indicated. It is emphasized that measurement tasks in conditions of information uncertainty caused by inaccuracy, incompleteness, fuzziness of data, their small volumes and uniqueness, incompleteness of models of the measuring object and the environment of its functioning, should be solved on the basis of regularizing methods that ensure the stability of the solutions obtained, as well as with the involvement of knowledge about the object and the environment, which makes it possible to compensate for the lack of completeness data. Therefore, in the article, to solve measurement problems in these conditions, it is proposed to use a regularization approach (RBP) and technologies based on it: Bayesian intelligent measurements (BII) and Bayesian intelligent technologies (BIT).Methodological principles and basic analytical dependencies of BII and BIT are given. The article proposes a methodology for creating a system for evaluating the quality of artificial intelligence (AI) solutions, in particular, evaluating the reliability of solutions, which determines the degree of confidence in the solutions obtained. The methodology is based on scaling of measuring solutions and implemented as a system of metrological support of AI solutions. Within the framework of this methodology and technologies of metrological support, complexes of metrological characteristics are proposed that determine accuracy, reliability (level of errors of the XNUMX-st and XNUMX-nd kind), which ensures traceability and transparency at each stage of obtaining solutions. Practical examples of the implementation of RBP methods and technologies for solving applied problems and creating intelligent measuring systems are given.
Keywords
Artificial intelligence, Metrology, Regularizing Bayesian approach
