In the context of the rapid development of information technologies, software quality is becoming critical for the successful operation of organizations in various industries. The growing complexity of modern software solutions requires the involvement of highly qualified specialists in software testing and quality assessment, capable of effectively identifying shortcomings and ensuring that the product meets established standards. At the same time, assessing the level of competence of such experts remains a difficult task, which is often based on subjective criteria and methods. The relevance of the study is due to the acute need of the modern IT market for objective tools for assessing the professional level of specialists, especially in the field of software quality assurance. Traditional approaches to qualification assessment, such as interviews, test tasks or resume analysis, often do not provide a complete and objective picture of the expert's competence. This problem becomes especially acute in the conditions of the global labor market, when companies are forced to evaluate specialists remotely, relying only on a limited set of data on their experience and skills. Today, software has become an integral part of many areas of our everyday life - from automation and optimization of production processes to creating comfort for an individual. The object of the study is the process of determining the level of competence of experts in software quality assessment. The subject of the study is a mathematical model for calculating the level of competence of an expert. The practical value of the results of the work is determined by the possibility of using the developed system by HR managers for effective selection of specialists, by heads of QA departments for the formation of balanced testing teams, by certification centers for objective assessment of competence, as well as by the experts themselves for planning their own professional development. Conclusion the developed mathematical model for calculating the level of competence of an expert allows you to reduce the time for assessing the competence of specialists, minimize the influence of subjective factors when making personnel decisions, and optimize the distribution of human resources in software development and testing projects.

Relevance. Modern artificial neural network models require significant energy and other resources for training and operation. Training generative models involves vast amounts of data. At the same time, these models face challenges related to the trustworthiness of the information they generate. An alternative to current paradigms of building and training neural networks is the development of energy-based models, which could potentially overcome these shortcomings and bring information processing closer to biologically and physically grounded processes. However, existing energy-based models differ little from classical models in terms of their limitations and drawbacks. Therefore, developing new approaches to modeling energy-based information processing in neural networks is highly relevant. The object of research is the process of information processing in artificial neural networks. The subject of the research is the mathematical models for the construction and training of artificial neural networks. The purpose of this paper is to develop and experimentally validate a theoretical framework that postulates the energetic nature of information and its role in the self-organization and evolution of complex information systems. Research Results. A fundamental theory is proposed, describing information as a structure of perceived external energy parameters that govern the processes of forming the internal energetic structure of a system—its model of the external world. This theory encompasses concepts of energy landscapes, principles of energy-based structural and parametric reduction, and a critical analysis of existing computational paradigms. Experimental studies on the construction and training of the developed energy-based model confirm its high generalization ability in one-pass training without using the backpropagation algorithm on ultra-small training datasets.

In the face of contemporary geopolitical challenges and transformative processes, particularly the decentralization reform, the objective assessment and forecasting of regional stability have become critical tasks for ensuring Ukraine's sustainable development. This research addresses the lack of comprehensive, automated tools for analyzing the condition of territorial communities by developing and validating an innovative model based on artificial intelligence. The methodological foundation of this work is the development and testing of a multi-task deep learning neural network designed to simultaneously solve four related tasks. The model concurrently performs two regression tasks to predict the precise numerical values of the Regional Security Index (RSI) and the Quality of Life Index (LI), as well as two classification tasks to determine the categorical levels of these indices (low, medium, high). The theoretical basis for the formation of these target indices is the Quadruple Helix concept, which describes the synergistic interaction between government, business, the scientific community, and civil society. The model was trained on a unique dataset covering 1469 Ukrainian territorial communities and containing heterogeneous socio-economic and security indicators. The experimental results demonstrated the high efficiency of the developed approach. On the test set, the classification accuracy reached 93.9% for the Regional Security Index and 85.0% for the Quality of Life Index. In the regression tasks, the model showed low mean absolute error values, indicating high predictive accuracy for both categorical levels and specific index values. The study concludes that the created model is a powerful and effective tool for monitoring, analyzing, and forecasting the dynamics of regional development in Ukraine. The results can be used by state and local government bodies to develop targeted policies aimed at enhancing the resilience, cohesion, and attractiveness of Ukrainian regions.

Public health emergencies demand fast, dependable analytics that combine real-time signals with trustworthy historical data. Open, interoperable platforms that support streaming and batch workflows can shorten the time from detection to action while preserving data quality and auditability. Aim: To design and justify an information system architecture for analyzing epidemic threats under emergency conditions that is scalable, reliable, and fit for integration with clinical and non-traditional data sources. Methods: We conducted a structured review of three data analytics architectures (Lambda, Kappa, Delta) and mapped their strengths and limits to crisis surveillance needs. Based on functional and non-functional requirements, we specified a Delta Lake–based lakehouse with bronze-silver-gold tiers, unified batch/stream ingestion with Spark Structured Streaming, ACID tables with time travel and schema control, and an analytics layer that supports forecasting with MLOps for monitoring, drift checks, retraining, and lineage. Results: The proposed architecture meets core emergency needs for timeliness, integrity, and reproducibility through ACID transactions, versioned datasets, and curated tiers; supports standards-based interoperability and the inclusion of wastewater, mobility, and other environmental feeds; provides a single code path for batch and streaming to reduce reconciliation burden; and sets operational guardrails for latency versus cost when running many near-real-time tables. We outline practical considerations for quality checks in the silver tier, promotion rules to gold, and model governance. Conclusions: A Delta-based lakehouse offers a clear path to an emergency-ready surveillance platform that scales with data growth, integrates heterogeneous sources, and supports reliable forecasting. The next steps are a pilot deployment with public health partners, live latency and cost measurements, and prospective validation of forecasting and alerting in real-world settings.

Research objective is to estimate the parameters of extreme value distribution laws constrained by parameters using the maximum likelihood method and to model the conditions of their occurrence with STATGRAPHICS and MATLAB tools. The subject of the study includes Anscombe's quartet and the first (Gumbel), second (Fréchet), and third (Weibull) laws of extreme value distribution. The research method involves numerical methods for solving systems of equations obtained by the maximum likelihood method, as well as statistical modeling techniques. The results of research show, through the example of statistical data analysis of Anscombe's quartet, the necessity of verifying the correspondence between the physical content of the studied processes and the applicability of extreme value distribution laws for their analysis. The linear regression equation, which corresponds to all possible combinations of this quartet’s data, is not optimal according to the criterion of maximum coefficient of determination. Using this criterion, it has been established that different data pairs have different nonlinear regression equations. Ignoring this fact may lead to errors in managing the processes they model. It has been shown that the data presented in Anscombe's quartet follow Gumbel's law, although the construction scheme of Anscombe's quartet does not correspond to the conditions of its occurrence. The limiting laws of extreme value distribution are presented in a form convenient for practical application in the design of engineering structures with constraints on the parameters of these distributions, caused by the specifics of the design. Parameter estites of extreme value distribution laws were performed using the maximum likelihood method. A numerical method for solving the corresponding systems of equations was considered. For modeling the scheme of extreme value occurrence, a matrix of random variables was generated, with twelve columns simulating monthly observations of a certain geophysical phenomenon, and one hundred rows representing a century of observations. The possible distributions of the elements of these matrices (initial distributions) were assumed as follows: double exponential distribution, Laplace distribution, lognormal distribution, Rayleigh distribution, normal distribution, Champernowne distribution, maximum extreme value distribution, minimum extreme value distribution, Birnbaum-Saunders distribution, Burr Type XII distribution, generalized extreme value distribution, inverse Gaussian distribution, Weibull distribution. The conditions of invariance of the extreme statistics distribution laws concerning the initial distributions were confirmed in only 12 out of 48 possible cases when modeled with STATGRAPHICS, and only in 5 out of 18 possible cases when modeled with MATLAB. The modeling results revealed a significant difference between the actual and theoretically possible extreme value distribution laws, which may be due to the peculiarities of the random number generation algorithms and the choice of the best-fit distributions used in STATGRAPHICS and MATLAB systems.

This study focuses on the automated process of determining sentiment (emotional coloring) in text messages from Telegram channels to enhance Ukraine’s information security. The principal challenge addressed lies in the need for rapid and accurate detection of negative, positive, or neutral messages within large-scale data streams without additional fine-tuning on local datasets. The essence of the results obtained is the implementation of a zero-shot classification approach, based on the multilingual transformer model XLM-RoBERTa, which in the experiment yielded the following metrics: Accuracy = 0.4718, Precision = 0.7138, Recall = 0.4718, and F1 Score = 0.5044. Owing to the model’s strong ability to generalize lexico-semantic patterns, a stable compromise between Precision and Recall was achieved, thereby increasing the efficiency of message analysis in large data volumes. These results are explained by the architectural features of XLM-RoBERTa, primarily its multilingual nature and deep layer structure, which enable proper handling of multilingual texts without dedicated local training. Conclusions. The proposed approach is advisable when there is a large, diverse corpus of data requiring prompt detection of potential negative informational influences and timely counteraction. Practically, this significantly reduces the time spent on manual monitoring of the information space and eases the burden on analysts, thereby strengthening the ability of organizations or information security units to respond rapidly to destructive content. The research results can also be integrated into decision support systems, serving as a foundation for the development of software solutions aimed at monitoring the information space.

The object of the study is the process recognition of pedestrian red light traffic. The subject of the study are the methods of process recognition of pedestrian red light traffic. The purpose of the paper is to improve the efficiency of real-time pedestrian red light traffic recognition model. The results obtained. The pedestrian red light traffic recognition model based on real-time object detection architecture YOLOv8 was proposed. The architecture and characteristics of YOLOv8 model, including its improved network structure, multi-scale detection ability, and adaptive anchor adjustment were introduced in detail. To demonstrate the efficiency and benefits of applying the YOLOv8 model, its performance was evaluated in various scenarios. Conclusions. Experiments have confirmed the efficiency of the proposed method. The use of the developed method based on the YOLOv8 architecture allowed to increase precision up to 0.935. Overall, the average performance across all categories is 0.851, which means that the model has a relatively high detection accuracy. In addition, model has a high-speed index.

In the article, the issue of mistaking birds for UAVs, which is considered one of the main problems in the process of detecting unmanned aerial vehicles (UAVs) through cameras, was considered. The object of the study is a method of training an artificial intelligence model so that the detection system does not mistake drones for birds. In order to solve this problem, as an alternative to creating a new complex model, the retraining method was used by making some changes to the previously developed model. The aim of the study is to train an artificial intelligence model using certain methods to detect drones and prevent them from being mistaken for birds. The main result of the research is the training of the artificial intelligence model with the application of the proposed methods for training the model and increasing the accuracy indicators. Also, a dataset was created for both classes, an algorithm was written to retrain the pre-trained model, a fine-tuning process was performed on the model to improve the training and validation performance, and the results were graphically expressed.

The object of the study is the process of lossy compression of noisy images and their post-filtering. The subject of the study is the approach to efficient two-stage processing (compression and post-filtering) for better portable graphics (BPG) coder and prediction of its efficiency. The goal of the study is to analyze performance characteristics of the considered two-stage approach and to propose an approach to their prediction. Methods used: numerical simulation, regression, statistical analysis. Results obtained: 1) the considered approach advantage is that it is able to provide improvement of quality of compressed noisy image under condition that an image is compressed with compression ratio smaller than that one corresponding to optimal operation point; 2) the approach efficiency depends on several factors including noise intensity, image complexity, and filter type and parameters; 3) the main characteristics of the two-step procedure can be quite accurately predicted in advance and this allows offering useful information for decision undertaking on what value of the coder parameter to apply; 4) this leads to either improving the compressed and processed image quality compared to its original version or, at least, to avoiding quality degradation. Conclusions: based on the results of the study, it is worth 1) predicting performance characteristics for the two-stage processing; 2) adapting the processing to image complexity and noise intensity.

Topicality. The concept of fog computing made it possible to transfer part of the data processing and storage tasks from the cloud to fog nodes to reduce latency. But in batch processing of integrated data streams from IoT sensors, it is sometimes necessary to distribute the tasks of the batch between the fog and cloud layers. For this, it is necessary to decompose the formed package. But the existing methods of decomposition do not meet the requirements for efficiency in high-density IoT systems. The subject of study in the article are methods of decomposition of integrated data streams. The purpose of the article is to develop a method of decomposition of an integrated data stream in a dense high-density Internet of Things fog environment. This will reduce the processing time of operational transactions. The following results were obtained. The concept of decomposition of integrated information flows in the foggy layer was implemented to transition from the batch mode to the flow mode of task processing. Within the framework of the concept, a method of selecting elementary task flows from an integrated flow is proposed. An algorithm for decomposition of the integrated flow of tasks is proposed. Conclusion. A comparison of the proposed method of processing information flows in the foggy environment of high-density IoT with the existing approach is carried out. The results of the comparison showed that the proposed method is more suitable for deployment in conditions of limited network and computing resources. It is advisable to use it on nodes of fog computing systems with a high density of IoT sensors.