Nataliia Lashchevska

Dr. Nataliia Lashchevska

orcid

Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
30 March 2023 in Visnyk NTUU KPI Seriia - Radiotekhnika Radioaparatobuduvannia
Verified
Journal Article
Open access
PDF file
ONO. Yu. Nedzelskyi
O. Yu. Nedzelskyi
NLNataliia Lashchevska
Nataliia Lashchevska
The main goal of the super resolution method is to create a higher resolution image from a lower resolution image. High-resolution images provide a high pixel density, hence more detail in the original image. The need for high resolution is widespread in computer vision techniques, pattern recognition applications, or general image analysis. However, high-resolution images are not always available. This is due to the fact that the conversion processes and processing methods require ultra-powerful processes, and the equipment for obtaining high-resolution images is expensive. These problems can be overcome by using image processing algorithms that are relatively inexpensive, which has led to the concept of super-resolution. This has the advantage that it can cost less and existing low-resolution imaging systems are readily available. High resolution is essential in medical imaging for diagnosis. Many applications require zooming into a specific image area, where high resolution becomes essential, such as surveillance, forensics, and satellite imaging. The method is presented in this paper, using a convolutional neural network to reproduce super-resolution images, directly performs the conversion from a low-resolution image to an image similar to the original. To speed up the output time, the proposed method performs most computational operations in low-resolution space, while reducing the sampling does not lead to information loss. The main task of the neural network is to reconstruct the distorted image and find the ideal reconstruction function, according to which, in fact, a neural network of a simple structure creates high-quality images with better performance, such as resolution, signal-to-noise ratio, with less time spent on image restoration. During the experiment, we determined an algorithm by which the proposed neural network can reconstruct any image with different types of distortion. The super-resolution method is implemented using the python 3.6 programming language and the tensorflow and tensorlayer software modules for convolutional neural networks. Graphical data of signal-to-noise ratio, structural similarity, and loss plots are obtained using the tensorboardX module.

19 October 2020
Verified
Book Chapter
Closed access
PDF file
AMAndrey Movchanyuk
Andrey Movchanyuk
OLOleksandr Luhovskyi
Oleksandr Luhovskyi
VFVolodymyr Fesich
Volodymyr Fesich
ISIryna Sushko
Iryna Sushko
NLNataliia Lashchevska
Nataliia Lashchevska

26 September 2025 in Cybersecurity: Education, Science, Technique
Verified
Journal Article
Open access
PDF file
ABAndrii Balvak
Andrii Balvak
NLNataliia Lashchevska
Nataliia Lashchevska
НЛНаталья Лащевська
Наталья Лащевська
The article presents a systematic review of contemporary scientific approaches to solving the Storage Location Assignment Problem (SLAP), managing the Order Picking Problem (OPP), and implementing Robotic Mobile Fulfillment Systems (RMFS) in warehouse logistics using machine learning methods. The study was based on a selection of publications from the Scopus database, compiled according to the PRISMA 2020 methodology, ensuring transparency and reproducibility of the analysis. As a result, 20 scientific publications were selected, each containing experimental and applied results related to logistics center automation. Task types, methods, and implementation technologies structure the literature analysis. It covers classical heuristic algorithms, intelligent systems, hybrid strategies, and machine learning-based approaches, including deep reinforcement learning, time series clustering, and associative analysis. Emphasis is placed on practical applications that aim to improve storage efficiency, reduce transportation costs, shorten order-picking times, and minimize physical strain on personnel. Particular attention is given to robotic systems that optimize movement routes and reduce overall operational time. The most promising directions include algorithms combining historical data analysis, product correlations, and preliminary demand classification. It was found that combined models integrating temporal features and associative relationships are effective in adaptive inventory placement and order batching. The results are systematized based on features such as task type, applied method, presence of machine learning, and implementation environment. The article also outlines future research directions, including adaptive clustering with seasonality considerations, predictive storage updates, Long Short-Term Memory (LSTM) model integration, and attention to energy-related aspects. As a result, the conducted research enables us to form a comprehensive understanding of the current state of scientific approaches in warehouse process optimization, providing the basis for the development of complex intelligent management systems that prioritize flexibility, efficiency, and the sustainable development of logistics centers.

2 June 2026 in Sustainability
Verified
Journal Article
Open access
PDF file
Viacheslav OsadchyiViacheslav Osadchyi
Viacheslav Osadchyi
ASAnton Shantyr
Anton Shantyr
OZOlha Zinchenko
Olha Zinchenko
Andrii BondarchukAndrii Bondarchuk
Andrii Bondarchuk
NLNataliia Lashchevska
Nataliia Lashchevska
Kateryna OsadchaKateryna Osadcha
Kateryna Osadcha
Artificial intelligence (AI)-related harms are increasingly attributed to governance failures rather than to isolated technical malfunctions. This article reframes AI governance as a core managerial competence grounded in leadership authority, accountability design, and organizational communication. The study addresses a persistent gap in higher education and managerial training, namely the insufficient preparation of future leaders to govern AI-mediated decision systems responsibly. Using a structured conceptual synthesis grounded in socio-technical systems theory and the organizational governance literature, the paper identifies recurring governance failure modes, including authority drift from human decision-makers to automated systems, diffusion of accountability, governance debt accumulation, and reliance on average-case performance metrics that obscure worst-case risks. To illustrate early governance readiness, an exploratory survey of senior university students—representing early-stage managerial cohorts—was conducted, resulting in the AI Governance Readiness Composite Score (AGRCS). The findings illustrate preliminary patterns in self-assessed governance readiness among early-stage managerial cohorts, without implying statistical generalization or population-level conclusions. The study does not seek statistical generalization but uses empirical signals to support conceptual arguments. The main contribution lies in positioning leadership authority, intervention capacity, and governance-related communication as central pillars of sustainable AI governance. The article translates these governance principles into an educational agenda, proposing sustainable pedagogy practices such as authority mapping, escalation rehearsals, worst-case simulations, and governance-focused learning environments. By framing AI governance as a leadership and communication challenge rather than a narrow technical problem, the study contributes to sustainable organizational development, responsible decision-making, and long-term societal trust aligned with the United Nations Sustainable Development Goals.

1 January 2025 in Cybersecurity: Education, Science, Technique
Verified
Journal Article
Open access
PDF file
ABAndrii Balvak
Andrii Balvak
NLNataliia Lashchevska
Nataliia Lashchevska
НЛНаталія Лащевська
Наталія Лащевська
The article considers the current scientific and practical problem of storage place assignment (SLAP) in e-commerce warehouses. Classical approaches, such as ABC analysis, optimise order picking routes while ignoring the operational costs of placing goods. This leads to an imbalance, resulting in the reduction of picking distances by increasing the labor intensity of the product put-away process. The analysis of recent studies has shown that despite the proven benefits of individual time series forecasting methods, the need for comprehensive methods remains relevant. In particular, there is a lack of approaches that would integrate the analysis of demand dynamics directly into solving the SLAP problem and balance the associated costs between the processes of placing and picking orders. The Time-Oriented Assignment of Storage Locations (TOASL) methodology is proposed, the fundamental principle of which is to combine the processes of order picking and goods placement within a single optimization problem. Its architecture consists of four key phases. The first stage involves data processing, which includes aggregation, filtering, and normalisation of demand time series. The modelling stage includes clustering of time series to identify groups of SKUs with similar dynamics. The simulation phase (creation of a virtual warehouse model and testing policies) and evaluation (calculation of key performance indicators and statistical verification) complete the cycle. The study, which covers the first two stages of the methodology, used a publicly available dataset from the Kaggle platform on transactions of an online electronics store. After data filtering, which selected only purchase events for target product categories, and aggregation by session, 471 unique SKU demand time series were generated. These series were normalized using the Min-Max method to focus on the shape and dynamics of demand, rather than on absolute volumes. A comparative analysis of three clustering algorithms was conducted: Ward's Agglomerative Hierarchical Clustering (AHC), k-means, and Self-Organizing Maps (SOM). The optimal number of clusters, which is 10, was determined by the elbow method. The AHC-Ward method demonstrated the best quality of partitioning, achieving the highest silhouette score with a value of 0.39. The performed grouping allowed to identify 10 sets of positions with similar dynamics. Based on the analysis of the 5 largest clusters, logically justified warehouse zoning principles were proposed, which provide for compact placement of goods with similar demand in adjacent aisles. It is shown that clustering of demand time series is an effective basis for further adaptive warehouse zoning.

1 January 2024 in Cybersecurity: Education, Science, Technique
Verified
Journal Article
Open access
PDF file
VSVladyslav Sosnovyy
Vladyslav Sosnovyy
NLNataliia Lashchevska
Nataliia Lashchevska
This article describes the problem of detecting malicious programs in running systems of users of mobile applications. Because users can download any application on their phone, which over time can pull up additional settings, which can store malicious routines for monitoring both personal life and their personal data, such as logins, passwords, bank data. The detection of such routines is based on dynamic analysis and is formulated as a weakly controlled problem. The article contains an analysis of information on the development of researchers who worked on detection models and methods such as: statistical and dynamic intrusion detection methods, anomaly detection model, settings classification methods, machine and deep learning methods. Machine learning, and especially deep learning, has become an extremely useful and interesting topic in cybersecurity over the past few years. In this context, the detection of malicious software has received considerable attention. The article considers the problem of detecting the activity of malicious software of mobile operating systems in the time domain by analyzing behavioral sequences of a large amount of industrial data. When malware executes on a system, its behavior consists of a series of distinct actions placed along the time axis, and there is only a subsequence of actions that lead to malicious activity. Very often, malicious software does not manifest itself immediately, and at some point in the execution, malicious activity is formed. Therefore, the main task and difficulty is to identify such a subsequence in the entire sequence of events. Due to this, it is proposed to develop a behavior model that would analyze the dynamic behavior of the program in the system during execution. For this, a sequence of API/function calls generated by the program at runtime is used as input data and a recurrent neural network (RNN) architecture is proposed to detect malicious activity. The article describes the training method of the proposed model and provides verification of its performance on a large sample of industrial data consisting of a large number of samples generated on the emulator farm. Many mobile phone vendors strive for hardware acceleration on the device to provide better support. Therefore, it can be considered that the deployment of a model based on RNM directly on the device as one of the security levels can become a viable solution. The test data of the model described in the article show sufficiently high positive results when detecting malicious activities.