S.S. Antsyferov – Dr. Sc. (Eng.), Professor, Moscow Technological University (MIREA)
K.N. Fazilova – Post-graduate Student, Moscow Technological University (MIREA)
I.V. Kuzmina – Ph.D.(Eng.), Associate Professor, Moscow Technological University (MIREA)
K.E. Rusanov – Ph. D. (Eng.), Associate Professor, Moscow Technological University (MIREA)
Nowadays, specialists of different fields of science pay attention to problem of creating intelligent systems, i.e. systems that process information at level of its «understanding» with a wide use of knowledge base. Urgency of this problem is caused both by growing in-formation complexity of phenomena and processes being investigated, and by increasing intensity of information flows being processed in a number of cases. Increase in requirements for accuracy (reliability) of results of processing only increases urgency of this problem.
The aim of this work is to establish building and functioning principles of intelligent information processing systems of isotropic spati-otemporal fields. As a mathematical model reflecting process of spatiotemporal field formation, use differential equation, according to which identification features of field realizations information are time-varying sets of spatially-oriented intensity gradients that form coherent configurations of different shapes and sizes, then realization measurability of information fields principle is possible. At the same time, there are a number of factors that have an interfering effect, which determines the probabilistic (stochastic) nature of access control and formation recognition procedure principles, allows to determine an acceptable set of field images. As a HIS knowledge model, a relational model based on theory of relations and deductive logic can be used as a basis for a system of products with output based on a given system of premises. Production form is used to represent knowledge in the form of implicit relations and connectives AND and OR between facts.
Carriers of products in proposed HIS are so-called active elements (AE), between which in process of adaptation (learning or self-learning), stochastic connections (each with each) are established, ensuring their information interaction and thereby solving a specific problem. The structure of AE includes the main knowledge base and a number of subsystems designed to extract knowledge, process external and internal information, form the goal, train and self-learn, control and diagnose, dialogue. Operation of AE included in the intelligence systems, there is a constant mutual information, which is a semantic operation that contains a set of interrelated procedures: generation, transmission, receiving, storing, perception, understanding, decision-making. Understanding semantic information (SI) active elements is defined through the semantic operation of comparing sign systems (external models) with the corresponding internal models – standards stored in AE memory and it reduces to formal logical operations. Decision-making – act of choosing one of a given number of alternatives according to established criterion, is reduced to formal logical operations and implemented by a software method or a special logic block. The presence of AE system, the a priori uncertainty of knowledge base, multivariance of the interdependent relationships determine the stochastic nature of hybrid intellectual systems functioning.
Principle of structurally stochastic approximation of spatiotemporal fields information and process of their processing is used as the basis for constructing and functioning of AIRS, allowing formation of field images in the process of adaptation and their subsequent recognition in solving specific problems. Dependence of adaptation rate on degree of mutual overlapping of the field images is investigated. Hybrid intellectual systems constructing principles that combine principles of structural-stochastic approximation with methods of knowledge processing representation are determined. GIS functioning phase portrait is established, which makes it possible to determine stable and unstable functioning.
- Antsyferov S.S. Forming the spectrum of thermal images of objects and recognize their images // J. Opt. Technol. 1999. V. 66. № 12. P. 44−48.
- Antsyferov S.S. Forming the spectrum of thermal images of objects and recognize their images // J. Opt. Technol. 1999. V. 66. № 12. P. 1047−1049.
- Antsyferov S.S., Evtikhiev N.N. Adaptation of information and recognition of biomedical systems // Biomedical Radioelectronics. 2001. V. 3. № 10. P. 52−57.
- Antsyferov S.S. Metrology virtual systems // Measurement technology. 2003. № 5. P. 17−21.
- Antsyferov S.S., Evtikhiev N.N. Adaptive pattern recognition system of spatiotemporal fields // Artificial Intelligence. 2004. № 3. P. 405−416.
- Antsyferov S.S., Evtikhiev N.N. Adaptive information processing of spatiotemporal isotropic fields // J. Opt. Technol. 2006. V. 3. № 10. P. 52−57.
- Antsyferov S.S. Metrology virtual systems // Artificial Intelligence. 2008. № 3. P. 18−27.
- Antsyferov S.S. Increase of the mental potential of adaptive information-distinguishing systems // Artificial Intelligence. 2010. № 4. P. 330−336.
- Antsyferov S.S. The general principles of construction and laws of functioning of intellectual systems // Artificial Intelligence. 2011. № 3. P. 6−15.
- Antsyferov S.S. Questions of metrological maintenance of intelligent systems // World of measurement. 2012. № 5. P. 46−51.
- Antsyferov S.S. Metrological basis of the information space in intelligent processing systems // Artificial Intelligence. 2012. № 4(58). P. 283−291.
- Antsyferov S.S., Rusanov K.E. Processing speed of intelligent systems // Artificial Intelligence. 2013. № 4(62). P. 259−265.
- Antsyferov S.S. Standardization of indicators of quality of products cognitive technologies // Science Intensive Technologies. 2014. V. 15. № 7. P. 7−13.
- Antsyferov S.S. Problems of Artificial Intelligence // Problems of Artificial Intelligence. 2015. № 0(1). P. 5−12.
- Denisov A.A. Teoreticheskie osnovy kibernetiki. Informacionnoe pole. L.: LPI. 1975.
- Antsyferov S.S., Fazilova K.N., Rusanov K.E. Indicators of non-equilibrium stability of cognitive systems // Problems of Artificial Intelligence. 2016. № 2(3). P. 4−11.
- Antsyferov S.S., Fazilova K.N., Rusanov K.E. Standardization of indicators of non-equilibrium stability of cognitive systems // Science Intensive Technologies. 2017. № 5. P. 15−20.
- Antsyferov S.S., Fazilova K.N., Rusanov K.E. Determination of indicators of stable functioning of neural-like systems // Materialy XXV Vseros. seminara «Nejroinformatika, eyo prilozheniya i analiz dannyh». 29 sentyabrya − 1 oktyabrya 2017 g. Krasnoyarsk: Institut vychislitel'nogo modelirovaniya SO RAN. 2017. P. 8−13.
- Antsyferov S.S., Fazilova K.N., Rusanov K.E. Simulated dynamic model of cognitive systems // Problems of Artificial Intelligence. 2017. № 2(5). P. 32−39.