Vol 27, No 9 (2021)

The paper describes a text comparison problem that arises during the update of open source programs in the presence of user changes. An algorithm based on calculating the Damerau — Levenshtein distance is used to correctly match the lines of two texts. The main difference between the proposed algorithm and the classical one is that the weights of such operations as insert, delete, replace, and permute are not constants, but depend on the values of individual text elements and their location. The paper also describes the functionality that allows the programmer to simplify the task of analyzing changes in program texts. The functionality is based on dynamic control of the comparison process, during which the programmer can manually establish correspondences between the compared texts. The effectiveness of the proposed method is checked when updating modules in the 1C:Enterprise system.

There is considered a problem of optimization methods comparing for the neural network regression task. Various optimization methods, such as stochastic gradient descent with momentum (SGDM), Adam and its modifications, AdamW and RAdam, were considered. To compare optimization method two regression task were formulated. Both tasks are connected with the preprocessing subtasks in the field of image analysis. The first considered task was the filtering blurred eye images on which confident recognition cannot be achieved. The training samples were generated by Gaussian blurring of the images. The blurring degree was estimated. The test and training sample for the assessment problem was formed on the basis of the BATH and CASSIA eye image databases. The second task was aligning faces in assessment image in face recognition systems. The training samples were generated by rotating face images, and rotation angle was estimated. To solve these tasks the direct estimation of parameters by solving the image regression problem by training neural network models is proposed. The adequate accuracy was acquired with all considered optimization methods for both tasks. Modifications of Adam algorithm show better results than original method. Both AdamW and RAdam methods reduced the error twice in comparison with Adam. The modification of the RAdam algorithm proposed in the work reduced the error by more than 1.5 times in comparison with the model trained by the original algorithm.

In this article, we propose detecting hidden elements in docx documents. Using the feature of the format structure, we analyze the most profitable stegocontainer. It is proposed to consider three options for investing in a stegocontainer in order to obtain the most objective estimates of the deviation. A comprehensive assessment is carried out from the tested data on the basis of which it was proposed to detect hidden elements in Microsoft Word documents.

The article suggests using elements described by a special potential function as elements of a neural network. The article describes the features and characteristics of a potential function that has an output signal close to a real physiological neuron. The description and structure of a neural network is given, the elements of which are described by a special potential function. To study the neural network and configure its elements, a mathematical apparatus of tensor calculus is proposed. The article describes the possibility of using the proposed neural network for recognizing and finding areas with specified parameters on a complex spatial surface. As an example, we describe the procedure for determining an area on a complex surface that has the specified geometric and qualitative characteristics.

An adaptive low-pass filter has been synthesized for automatic patterns detection of arbitrary mental movements in records of multidimensional electroencephalograms (EEG). The filter is based on the multivariate singular spectrum analysis. The effective bandwidth of the filter corresponds to the spectrum of the sought patterns of mental activity on the observed EEG time interval. The use of the synthesized filter provided a reliable automatic search for patterns and the correct determination of their time boundaries. The correctness of the results has been confirmed in experiments with 24 volunteers.

The paper discusses the effectiveness of digital technologies in the educational process. The authors consider the use of digital services in the university educational process in the context of the need to full transfer basic education into online format.

A comparative analysis of public and corporate digital services is carried out. Their advantages and disadvantages are considered the application effectiveness in the educational process. The article examines the following digital services: Zoom, Jitsi, MS Teams, Google Documents, Google Classroom, LMS HSE, Smart LMS. The authors propose the methods of organizing the e ducational process, involving the integration of these services.

The results of the students’ survey revealed the following features that affect the interaction of teachers and students: the Internet speed and availability, preferences in the ways of course materials obtaining, the desired forms of interaction between students and the teacher.

The paper describes the effective methods of conducting online classes in various conditions: group discussion of the prevoiced and cloud-uploaded presentations in chat, conducting classes in the form of a video-chat, using pre-prepared dynamic graphic illustrations, using gamification elements, laboratory work remote running on a virtual equipment, splitting the lecture into thematic sections and frequent changes in the students’ activity. The author s offer various online forms and methods of learning outcomes intermediate control for the disciplines of the university curriculum. This study offers solutions to the problems of organizing online classes, including in conditions of limited Internet traffic.