РУС ENG

№ 3(30)'2023

About

News
Aims and Scope
Founder and Publisher
Editorial board
Licensing terms
Privacy Statement
Plagiarism policy
Publication ethics
Archiving Policy
Subscription


For Authors

Instructions for authors
Reviewing proccess
Copyright Notice
License agreement
Article processing charges


Archive

All issues
Search


Contacts

Contacts


Section 1
Informatics, Computer Engineering and Control


Kharlamov Alexander, Samaev Eugeny, Kuznetsov Dmitry, Pantiukhin Dmitry.
SEMANTIC TEXT ANALYSIS USING ARTIFICIAL NEURAL NETWORKS BASED ON NEURAL-LIKE ELEMENTS WITH TEMPORAL SIGNAL SUMMATION

UDC 528.013
DOI 10.34757/2413-7383.2023.30.3.001
Language: English
Annotation:Text as an image is analyzed in the human visual analyzer. In this case, the image is scanned along the points of the greatest informativity, which are the inflections of the contours of the equitextural areas, into which the image is roughly divided. In the case of text analysis, individual characters of the alphabet are analyzed in this wayNext, the text is analyzed as repetitive language elements of varying complexity. Dictionaries of level-forming elements of varying complexity are formed, the top of which is the level of acceptable com-patibility of the root stems of words (names) in sentences of the text, that is, the semantic level. The level of semantics represented by pairs of root stems is virtually a homo-geneous directed semantic network. Re-ranking the weights of the network vertices corresponding to the root stems of individual names, as occurs in the hippocampus, makes it possible to move from the frequency characteristics of the network to their semantic weights. Such networks can be used to analyze texts that represent them: one can compare them with each other, c lassify and use to identify the most significant parts of texts (generate abstracts of texts), etc.
Keywords: text analysis; language model; neural network; transformer model; semantic analysis of texts; artificial neural networks based on neurons with temporal summation of signals; language levels; semantic level; TextAnalyst technology for semantic text analysis; applications.

S. S. Antsyferov, K. N. Fazilova, K. E. Rusanov.
«DOCUMENTARY INFORMATION FLOW» SYSTEMS STRUCTURAL CONSTRUCTION PRINCIPLES

UDC 681.518.9; 621.384.3
DOI 10.34757/2413-7383.2023.30.3.002
Language: Russian
Annotation:The article defines the principles of construction and functioning of the system of documentary information flows. The main stages of the development of the scientific direction, which are reflected in the documentary information flow, are presented. It is pointed out that the principles of the construction of the "documentary information flow" system are interrelated with the nature of the stages of the development of the scientific direction and that the principles of the structural construction of the "documentary information flow" system can be used in the creation of expert systems.
Keywords: documentary information flows, information base, intelligent control systems, input data, semantic template blocks, information, control.

A. A. Nikitina, S. I. Ulanov.
DETECTION OF OBJECTS ON THE GROUND BY INTELLIGENT ROBOTS IN A RAPIDLY CHANGING ENVIRONMENT

UDC 528.013
DOI 10.34757/2413-7383.2023.30.3.003
Language: Russian
Annotation:The article presents several of the most effective approaches for solving the problem of detecting objects on the ground by intelligent robots in a rapidly changing environment. The application of computer vision, machine learning based on self-learning convolutional neural network is justified. The study of this problem has shown that the most effective result will be the use of these approaches in a complex.
Keywords: self-learning convolutional neural network, FPGA, GPS data, data reduction.

Ya. S. Pikalyov, T. V. Yermolenko.
ABOUT NEURAL ARCHITECTURES OF FEATURE EXTRACTION FOR THE PROBLEM OF OBJECT RECOGNITION ON DEVICES WITH LIMITED COMPUTING POWER

UDC 004.932.72
DOI 10.34757/2413-7383.2023.30.3.004
Language: Russian
Annotation:This work is devoted to the study of the effectiveness of various neural network models in the tasks of object detection and classification on devices with limited computing power. The authors use a two-step approach based on the Faster R-CNN architecture to detect an object in an image and recognize it. The basic network is the main block in the Faster R-CNN structure that affects the quality and performance of the entire system. The paper presents the results of numerical studies of the effectiveness of various network architectures according to criteria such as the separating ability of high-level features, clas-sification accuracy, the amount of RAM occupied, computational complexity. An integral assessment of the effectiveness of the models is proposed, taking into account the above criteria. The best value according to the integral criterion was shown by the hybrid network EdgeNeXt-S, which indicates a good balance of this model between performance, robustness and accuracy in computer systems
Keywords: computer vision, object detection, backbone networks, deep learning, clusterisation, edge devices.

Y. Y. Potreba, N. E. Gubenko.
ANALYSIS OF METHODS AND MEANS OF PREVENTING CONFIDENTIAL DATA LEAKS

UDC 004.056.5
DOI 10.34757/2413-7383.2023.30.3.005
Language: Russian
Annotation:The article analyzes methods and means of protecting confidential information. The search for the most effective approaches to ensure the security of confidential data has been carried out. The advantages and disadvantages of the selected approaches are described. The functions of the data leak prevention system are classified.
Keywords: information security, confidentiality, data leakage.

Section 2
Math modeling


I. I. Maksimenko.
IDENTIFICATION OF ALGEBRAIC OBJECTS WITH A BASED ON BAER METRIZATION OF OBJECTS CLASS

UDC 519.713.4
DOI 10.34757/2413-7383.2023.30.3.006
Language: Russian
Annotation:The article discusses the problem of identifying objects of a class with a standard on the basis of various mathematical structures (finite automata, unstructured sets, lattices, closed semirings) based on the introduction of the concept of representation and the Baire metric of a special type. A criterion for the existence of representations in terms of the properties of limit objects of a class has been found, which generalizes the previously found criterion for Mealy automata. For a finite definition of classes the criterion is constructive. This criterion indicates the presence of a deep connection between the process of identification with the standard and the properties of the limiting points of the metric space of a class of objects.
Keywords: identification, representation, metric, fragment, cofragment, limits objects.

K. B. Fam, P. M. Murashev, V. N. Bogatikov.
FUZZY MODEL OF DIAGNOSTICS OF TECHNOLOGICAL PROCESSES

UDC 519.4
DOI 10.34757/2413-7383.2023.30.3.007
Language: Russian
Annotation:In this paper, a system for ensuring the safety of the technological process using an algorithm for determining the degree of safety at the current moment is presented. To represent an ordered set of faults, the Hasse diagram construction algorithm was used.
Keywords: diagnostics of technological processes, Hasse diagram, fuzzy equality, fuzzy inclusion, graph search.