This article is devoted to a comparative analysis of the resilience of ResNet18 and ResNet50 neural networks to adversarial attacks on training sets. The issue of the importance of ensuring the safety of learning sets is considered, taking into account the growing scope of artificial intelligence applications. The process of conducting an adversarial attack is described using the example of an animal recognition task. The results of two experiments are analyzed. The purpose of the first experiment was to identify the dependence of the number of epochs required for the successful execution of an adversarial attack on the training set on the neural network version of the ResNet architecture using the example of ResNet18 and ResNet50. The purpose of the second experiment was to get an answer to the question: how successful are attacks on one neural network using modified images of the second neural network. An analysis of the experimental results showed that ResNet50 is more resistant to competitive attacks, but further improvement is still necessary.

Keywords: artificial intelligence, computer vision, Reset, ResNet18, ResNet50, adversarial attacks, learning set, learning set security, neural networks, comparative analysis

The article presents an analysis of the application of the Socratic method for selecting machine learning models in corporate information systems. The study aims to explore the potential of utilizing the modular architecture of Socratic Models for integrating pretrained models without the need for additional training. The methodology relies on linguistic interactions between modules, enabling the combination of data from various domains, including text, images, and audio, to address multimodal tasks. The results demonstrate that the proposed approach holds significant potential for optimizing model selection, accelerating decision-making processes, and reducing the costs associated with implementing artificial intelligence in corporate environments.

Keywords: Socratic method, machine learning, corporate information systems, multimodal data, linguistic interaction, business process optimization, artificial intelligence

The article examines the modular structure of interactions between various models based on the Socratic dialogue. The research aims to explore the possibilities of synthesizing neural networks and system analysis using Socratic methods for managing corporate IT projects. The application of these methods enables the integration of knowledge stored in pre – trained models without additional training, facilitating the resolution of complex management tasks. The research methodology is based on analyzing the capabilities of multimodal models, their integration through linguistic interactions, and system analysis of key aspects of IT project management. The results include the development of a structured framework for selecting suitable models and generating recommendations, thereby improving the efficiency of project management in corporate environments. The scientific significance of the study lies in the integration of modern artificial intelligence approaches to implement system analysis using multi – agent solutions.

Keywords: neural networks, system analysis, Socratic method, corporate IT projects, multimodal models, project management

In this paper, an analysis of the calculation results is carried out, which makes it possible to assess the real impact of impact and explosive effects from UAVs on the strength of reinforced concrete structures. Load limits are set, depending on the four most common types of walls. The previously published classification made it possible to identify the main parameters necessary for a detailed load calculation.

Keywords: extreme loads, explosion, self-supporting walls, reinforced walls, non-reinforced walls, load-bearing walls, UAVs, strength testing, building structures, shock waves

The paper considers the solution of an urgent scientific and technical problem of developing a tilt angle sensor designed for automated position measurement of optical and optoelectronic systems. The principle of operation of the tilt angle sensor or inclinometer is to measure the direction of acceleration of gravity in the coordinate system associated with the target object, which allows calculating the angles of deviation of the vertical axis of the object from the ideal position. Capacitive sensors measure the angle of inclination by monitoring changes in capacitance due to external influences and often have a fairly simple design. In this paper, we consider a sensor design that uses a metal ball moving in a dielectric tube with electrodes attached to it. To investigate the possibilities of optimizing the device, a mock-up was developed on which the change in capacitance was analyzed depending on the size of the ball, the diameter of the tube and the shape of the electrodes. An important goal of this study was to determine the optimal diameter of the ball inside the dielectric tube to obtain the most noticeable changes in capacitance. The best results were obtained using a metal ball with a diameter slightly smaller than that of a dielectric tube. To measure the capacitance the "capacity–frequency" conversion can be used, followed by measuring the latter using a microcontroller. The resulting capacitance values are very small, which leads to a significant impact of any connecting wires and requires reducing the distance between the primary converter and the signal processing circuit, which can be achieved using micromechanics and microelectromechanics technologies and integrating the entire sensor design into one miniature housing. At the same time, with a decrease in the size of the elements of the primary converter, the value of the detected capacitance will obviously decrease, which will significantly complicate its conversion into an electrical signal.

Keywords: tilt angle, electronic sensor, inclinometer, position control, capacity change, primary converter, auto generator, miniaturization

The article explores the actor model as implemented in the Elixir programming language, which builds upon the principles of the Erlang language. The actor model is an approach to parallel programming where independent entities, called actors, communicate with each other through asynchronous messages. The article details the main concepts of Elixir, such as comparison with a sample, data immutability, types and collections, and mechanisms for working with the actors. Special attention is paid to the practical aspects of creating and managing actors, their interaction and maintenance. This article will be valuable for researchers and developers interested in parallel programming and functional programming languages.

Keywords: actor model, elixir, parallel programming, pattern matching, data immutability, processes, messages, mailbox, state, recursion, asynchrony, distributed systems, functional programming, fault tolerance, scalability

The analysis of the environmental impact of the largest enterprises located in the Southern and Northern industrial zones of the linear city of Volgograd has been carried out, and the need to change approaches to designing a comfortable urban environment, which currently take into account the average data for characterizing the ecological state of a particular territory, has been shown. The analysis confirmed the need to take into account the local impact of industrial enterprises on the components of the urban environment when justifying the selection and planning of appropriate modern spaces within the framework of the Federal Project "Creating a comfortable Urban environment".

Keywords: urban environment, comfort, urban environment quality index, modern spaces, environmental analysis, environmental friendliness, environmental safety

The article contains the results of the inspection of power transmission line structures within the framework of the preparation of design documentation for the reconstruction of cable racks and power transmission line supports due to the increase in the diameter of cables and, as a consequence, the increase in the load on the racks located in the Chaunsky District of the Chukotka Autonomous Okrug. The article presents a geodetic survey of load-bearing structures (determination of deflections and displacements of structures), lists of defects and damages, verification calculations of structures, and measures to strengthen structures.

Keywords: bearing capacity, inspection of structures, overhead power lines, wooden structures, supports, overpasses

The article presents the existing methods of reducing the dimensionality of data for teaching machine models of natural language. The concepts of text vectorization and word-form embedding are introduced. The task of text classification is being formed. The stages of classifier training are being formed. A classifying neural network is being designed. A series of experiments is being conducted to determine the effect of reducing the dimension of word-form embeddings on the quality of text classification. The results of evaluating the work of trained classifiers are compared.

Keywords: natural language processing, vectorization, word-form embedding, text classification, data dimensionality reduction, classifier

The paper considers the issue of using a computer vision system to control the quality of products in the control algorithm of a mechatronic sorting station. Shoe products are chosen as an example. The developed system is based on machine learning methods for image recognition by segmentation. As a result, a neural network model was created, and a program was written for identifying and selecting objects using a camera for subsequent sorting of defective products. The program contains three modules: initialization for declaring all variables, models, classes, video stream from the camera; the main module, containing an internal loop for each segmented object; a subroutine for completing the work. The introduction of computer vision into the control algorithm increases the efficiency and flexibility of the quality control system, and improves the accuracy of measuring the parameters of objects for their subsequent sorting.

Keywords: mechatronic station, sorting, computer vision, image segmentation, neural network training, control algorithm

The application of a mechanical-empirical method of designing road pavement for predicting track formation on the surface of road structures is discussed in the article . The occurrence of ruts is one of the most common types of defects in road surfaces, significantly reducing safety and comfort of movement. MEPDG calculation models were used in this work for assessing the stability of a road structure to the accumulation of plastic deformations. Models takes into account factors such as the intensity and axial load of transport, climatic conditions, and the physic and mechanical properties of coating materials. An example of predicting the depth of a track using the considered method is given.

Keywords: rutting, mechanical and empirical method, road pavement, forecasting, MEPDG, traffic loads, climatic conditions

The article discusses the process of formation of cavitation bubbles. One of the effective designs of regulating equipment is an axial type valve, for which a mathematical description of the process of formation of cavitation bubbles has been developed. This description allows you to evaluate the bubble structure depending on the main operating and design dimensions of the valve.

Keywords: valve, cavitation bubbles, probability, functions

In the work describes the extreme filtering method and the author's approaches that allow adapting it to work in real time: frame-by-frame processing and the method with signal loading. Further, solutions are presented that can be used to implement the above on real devices. The first solution is to use the Multiprocessing library for the Python language. The second approach involves creating a client-server application and sending asynchronous POST requests to implement the frame-by-frame signal processing method. The third method is also associated with the development of a client-server application, but with the WebSocket protocol, not HTTP, as in the previous approach. Then, the results are presented, and conclusions are made about the suitability of the author's approaches and solutions for working on real devices. It is noted that the solution based on the use of the WebSocket protocol is of particular interest. This solution is suitable for both the frame-by-frame signal processing method and the method with value loading. It is also noted that all approaches proposed by the author are workable, which is confirmed by the time values ​​and the coincidence of the graphs.

Keywords: extreme filtering, frame-by-frame signal processing method, method with value loading, Multiprocessing, HTTP, WebSocket, REST, JSON, Python, microcontrollers, single-board computers

The article proposes an approach to automate the detection of polishing defects in blades using luminescent testing (LUM). Instead of manual visual inspection, a system was developed that utilizes a deep learning neural network for defect segmentation on images and a robotic setup for precise positioning of the camera and the blank. This ensures the repeatability of the inspection. The relevance is driven by the industry's need for high-precision and reliable real-time quality control methods. The mathematical model of the process, software architecture, hardware components, and the data collection process for neural network training are described. The results of applying the system for defect detection are presented. The development optimizes polishing processes.

Keywords: industrial blade polishing, intelligent video analytics, robotic optical scheme, mathematical model of technological process, Lum control

The article is devoted to the automation of the process of managing road construction works at a manufacturing enterprise. Among the means of communication in Russia, highways are in the first place in terms of length. Construction of new roads, repair and bringing the existing roads to regulatory requirements is a complex process that can be characterized as a project. The process of project-oriented management of road construction works is formalized, project limitations are defined. The enlarged milestones of project-oriented management of road construction works are highlighted, including the stages of initialization and implementation. The categories of system users and their functions are defined. A class diagram of the information system for managing road construction works is provided. An algorithm for the operation of an automated system for managing road construction works based on a project-oriented approach is developed and described in detail. Formalization of the calculation of the percentage of project readiness is carried out based on the significance coefficient. Examples of implementing the algorithm stages in the information system and generating analytical reports in the system are given. The reports generated in the system are described in detail. The economic efficiency of the proposed automation system is substantiated.

Keywords: road construction works, project-oriented management, highway, automation, reporting, significance coefficient, project, project resources, performance indicator, construction, repair