В статье представлена задача применения многоагентных систем при реализации концепции нулевого доверия. Обсуждаются возможные архитектуры многоагентных систем, используемых для решения задач сбора информации о состоянии системы, адаптации схем размещения точек применения политик безопасности. Рассматриваются особенности применения нейронной LSTM-се в системе анализа сетевого трафика на уровне координации многоагентной системы. The paper considers the problem of applying multi-agent systems in the implementation of the concept of Zero Trust. Possible architectures of multi-agent systems used to solve the problems of collecting information about the state of the system, adapting the rules of the schemes for placing points of application of security policies are discussed. The features of the LSTM neural network application in the network traffic analysis system at the coordination level of a multi-agent system are considered.

As part of this research, software for multimodal biometric authentication using neural networks is developed to improve the efficiency of information system user authorization. The architectures of artificial neural networks, which are involved in the processes of recognizing a person by facial image and voice, are given. The internationally used databases (DataSet) of images and audio recordings for training of neural networks are considered. The process of training neural networks, the formed database of biometric personal data and the results obtained by the authors are described.

Peculiarities of the automated realization of hashing biometric personal data, information resources of a user, and the information system in general are considered. The basis of the created mathematical model is an artificial neural network intended for hashing biometric images in the account administrator folder with the use of mathematical methods. The object of investigation is information systems. The subject of investigation is the means of information security with the use of biometric multifactor authentication.

Information security tools are an integral part of system users. The concept of information security implies the development and expansion of the scope of innovative technologies in information processing. To keep the information security system up to date, it is necessary to periodically update and supplement the structure of information protection, information security threat model and hardware and software complex. This paper analyzes the existing methods of information protection and proposes the implementation software system modules for personality recognition by photo and video, voice recognition, deep neural network and the creation of configuration for its weights file. Based on the generated data set, a method for synthesizing the parameters of a mathematical model of a convolutional neural network, presented as an array of real numbers, which are unique identifiers of a personal computer user, has been developed and proposed. This study uses the features of simulation modeling of user authorization systems, as well as the error function when compiling a neural network model. The training model of multi-factor biometric authentication is trained using categorical cross-entropy. The training sample is generated by adding distorted images from the database by changing the receptive fields of the convolutional neural network. The objective of this study is the application of new methods and means of protecting information of workstations from information threats. The result of the study is the developed information security system designed to ensure the information security of users of personal computers and workstations of enterprises.

Аутентификация относится к классическим средствам управления информационной безопасностью компьютерных систем предприятия, от качества которой зависит безопасность информационной системы. В данной статье описана процедура аутентификации пользователей информационной системы по изображению лица. Разработана архитектура искусственной нейронной сети, сформированы наборы биометрических персональных данных и проведено обучение на основе распознавания пользователей информационной системы по изображению лица. В рамках данного исследования проведена оценка функциональности архитектуры искусственной нейронной сети на международных банках данных (Dataset). При распознавании пользователей информационной системы по изображению лица были извлечены такие дескрипторы, как локальные бинарные шаблоны (LBP) и гистограмма ориентированных градиентов (HOG). Скомпилирована модель обучения нейронной сети на основе категориальной кросс-энтропии, сформирована конфигурация компиляционной модели (размер мини-выборки, количество эпох, функция активации, функция оптимизации). Разработанный программный модуль производит аутентификацию пользователей информационной системы по принципу «свой-чужой». Применение данных дескрипторов изображения позволяет повысить точность распознавания пользователей информационной системы (accuracy) и снизить значение функции потерь (loss). Реализован программный код системы мультимодальной биометрической аутентификации. Для оценки эффективности работы программного модуля приведены показатели ошибок первого и второго рода. Authentication belongs to the classical means of information security management of enterprise computer systems, the quality of which determines the security of the information system. This paper describes the authentication procedure of information system users by facial image. The architecture of an artificial neural network has been developed, biometric personal data sets have been formed and trained based on the recognition of information system users by facial image. As part of this research, the functionality of the artificial neural network architecture has been evaluated using international data banks (Dataset). Descriptors such as Local Binary Patterns (LBP) and Histogram of Oriented Gradients (HOG) were extracted when recognizing information system users by facial image. A neural network-training model based on categorical cross-entropy was compiled, and the configuration of the compilation model (mini-sample size, number of epochs, activation function, and optimization function) was generated. The developed software module authenticates users of the information system on “friend-or-for” basis. The use of these image descriptors allows increasing the accuracy of user authentication in the information system (accuracy) and reducing the value of loss function (loss). The program code of the multimodal biometric authentication system has been implemented. To assess the efficiency of the software module, the first and second type error rates are given.

Рассмотрены особенности автоматизированной реализации хэширования биометрических персональных данных, информационных ресурсов пользователя, информационной системы в целом. Основой созданной математической модели является искусственная нейронная сеть, предназначенная для хэширования биометрических образов в папке администратора учётной записи с применением математических методов. Объект исследования - информационные системы. Предмет исследования - средства защиты информации с использованием биометрической многофакторной аутентификации. This article describes features of automated implementation encryption of both biometric data and user information resources as a whole. The created mathematical model is based on an artificial neural network designed for encryption of biometric images in the account administrator folder using mathematical methods. The object of the study is the information systems of the PC. The subject of research is the tools of protection of user information using biometric multifactor authentication.

The article describes the development and integrated implementation of software modules of photo and video identification system, the system of user voice recognition by 12 parameters, neural network weights, Euclidean distance comparison of real numbers of arrays. The user's biometric data is encrypted and stored in the target folder. Based on the generated data set was developed and proposed a method for synthesizing the parameters of the mathematical model of convolutional neural network represented in the form of an array of real numbers, which are unique identifiers of the user of a personal computer. The training of the training model of multifactor authentication is implemented using categorical cross-entropy. The training sample is generated by adding distorted images by changing the receptive fields of the convolutional neural network. The authors have studied and applied features of simulation modeling of user authorization systems. The main goal of the study is to provide the necessary level of security of user accounts of personal devices. The task of this study is the software implementation of the synthesis of the mathematical model and the training neural network, necessary to provide the maximum level of protection of the user operating system of the device. The result of the research is the developed mathematical model of the software complex of multifactor authentication using biometric technologies, available for users of personal computers and automated workplaces of enterprises.

Kinetic and mathematical models are given in the paper for the polymerization process under consideration. Mathematical interpretations of direct and inverse kinetic problems are given. The latter is reduced to finding the values of the kinetic constants of the elementary stages of the polymerization process. For preliminary localization of their values and narrowing of the permissible range of their variation, the paper proposes a method for constructing a functional surface in the basic space.

The purpose of this work is to automation of the informativity content analysis of kinetic parameters based onde composing a chemical reaction independent route. The software for determining the basis of nonlinear parametric functions of kinetic measurements for the mechanisms of complex chemical reactions is developed and described.

A graph-theoretical algorithm based on the depth-first search (DFS) procedure of graph analysis is proposed for determining a complete set of homodesmic reactions (HDRs) of organic compounds. The complete set of HDRs is constructed decomposing the molecular graph of a compound into a series of simpler components. Based on the resulting DFS algorithm, a program for plotting the HDR basis is devised that includes automated generation of the structures of homodesmic reaction participants and primary thermochemical analysis (i.e., calculating the heat effects of HDRs from absolute enthalpies of reactants obtained via user-selected quantum chemistry). The work of the algorithm is illustrated by the example of vinylcyclopropane. The use of complete sets of HDRs for the thermochemical analysis of molecular energetics is studied to determine its advantages: high precision; the possibility of controlling the reproducibility of data and screening out inconsistent or erroneous thermochemical data; quantitative consideration of the nonvalence structural effects in organic compounds; and block analysis of the thermochemistry of structurally similar organic compounds, which is analogous to the familiar concept of active thermochemical tables.

The work contains a mathematical description of diene polymerization with a neodym-based Ziegler-Natta catalysts. The resulting system of differential equations was reduced from infinite to finite by the method of moments. For this, the basis of nonlinear parametric functions was searched with a method allowing obtain expressions that reflect the kinetics of the polymerization process. The obtained basis of two functions enables us to construct a surface containing local regions of the minimum of the objective function which define the starting points for search of optimum kinetic constant values.

In the work of solving the problem of preliminary search for the optimal values of the kinetic rate constants of the elementary stages of polymerization processes. A statement of the problem and an algorithm for its solution in three steps are given. Graphical images characterize the local areas of finding the optimal values of the constants. In this area, one critical area remains, in another place, however, this period remains with the researcher.

Предлагаемый в настоящей работе теоретико-графовый алгоритм для анализа энергетики химических соединений основан на гомодесмическом подходе. В работе описаны основные этапы конструирования базисного набора гомодесмических реакций. Работа алгоритма проиллюстрирована на молекуле метилциклобутан. В разработанной программе для построения гомодесмических реакций использовался метод поиска в глубину DFS. Application of the DFS Algorithm to the Homodesmic Analysis of Chemical Compounds The graph-theoretic algorithm proposed in this work for the analysis of the energetics of chemical compounds is based on the homodesmic approach. The paper describes the main stages of constructing a basic set of homodesmic reactions. The operation of the algorithm is illustrated on the methylcyclobutane molecule. In the developed program for constructing homodesmic reactions, the DFS depth-first search method was used.

The subject of this study is the inverse problems of identifying the mechanisms of complex chemical reactions. A graph-theoretic method for determining the basis of the parametric functions of the kinetic parameters of mathematical models for complex catalytic reactions is proposed. Based on the developed algorithm, a program for analyzing the information content of a mathematical model of the mechanism of catalytic reactions has been implemented. The operation of the algorithm is illustrated by the reaction of isotopic exchange of protium for deuterium.

Mathematical modeling of catalytic processes is necessary for the complete and accurate description, as well as for controlling the quality and physicochemical studied of catalysts. In the paper, theoretical issues of industrial catalysis are discussed. The work is devoted to theoretical graph analysis of informativity of kinetic parameters of the model of a complex chemical reaction. The aim is the development and automation of algorithm for determining basis of nonlinear parameter functions in solving inverse problems of chemical kinetics in order to define the number and form of independent combinations of rate constants of elementary stages. A program package for analysis of informativity of kinetic parameters of the mathematical model of a complex catalytic reaction is developed and described. The obtained functional relations between the kinetic parameters can be useful for experimentalists in physicochemical interpretation and analysis of mechanisms of chemical reactions. In other words, the proposed method allows independent combinations of kinetic constants to be distinguished that results in shortening the number of the model parameters and, as a consequence, enhance the accuracy of the mathematical model. The mechanism of hydrogen oxidation over a platinum catalyst is given as an example of the use of the software.