Перегляд за Автор "Shevchuk, P."

Зараз показуємо 1 - 3 з 3
  • Ескіз
    Документ
    Deep learning neural network architecture for determining sunflower growth stage from visual data
    (2025) Malaydakh, V.; Molchanova, M.; Shevchuk, P.; Mazurets, O.; Мазурець, Олександр Вікторович
    This paper presents the development of a deep learning neural network architecture based on EfficientNet-B0 for determining the growth stage of sunflower plants from visual data. The research addresses the need for automated, accurate, and resource-efficient phenological analysis in precision agriculture. The proposed model processes 224×224-pixel images and classifies sunflower development into eight phenological stages, achieving a classification accuracy of approximately 92% with minimal computational overhead. Key architectural features include MBConv blocks, compound scaling, and optimised inference suitable for mobile and UAV deployment. The system enables real-time monitoring, early detection of anomalies, and adaptive agronomic decision-making. Results highlight the potential for integrating this lightweight architecture into edge-computing platforms to enhance sustainability and efficiency in large-scale agricultural operations
  • Ескіз
    Документ
    Software architecture of information system for exchanging LLM thematic prompts
    (2025) Denysenko, B.; Shevchuk, P.; Molchanova, M.; Mazurets, O.; Мазурець, Олександр Вікторович
    The article presents the design of a software architecture for an information system focused on the exchange of thematic prompts for large language models. As the use of LLMs grows across various domains, the need for a structured platform to manage, share, and evaluate prompts becomes critical for productivity, reproducibility, and collaboration. The system is based on the MVC architectural pattern (Laravel framework) and includes key modules for authentication, prompt management, subscription/payment handling, and administration. The proposed solution enables role-based access, prompt versioning, user feedback, and integration with LLM APIs, laying the foundation for scalable, collaborative, and transparent prompt engineering
  • Ескіз
    Документ
    Software for Text Messages Reliability Analysis Based on the Machine Learning Models Ensemble
    (2024) Shevchuk, P.; Molchanova, M.; Mazurets, O.; Мазурець, Олександр Вікторович
    The developed model is the closest to predicting the data on the labeled set, showing 4 errors out of 50 test samples, which is 92% accurate when analyzing the reliability of text messages. The implementation of the test software in the form of a website was carried out by integrating Scikit-Learn and Flask technology. It is proposed to add four different machine learning models to the ensemble — logistic regression, decision trees, gradient boosting, and random forest, on the basis of which a weighted estimate of the credibility of text messages will be formed, which is calculated as the sum of the influence coefficients of each model multiplied by the output of the corresponding classifier model.