Kushnerov Oleksandr

Mr. Kushnerov Oleksandr

Sumy State University, Sumy, Sums’ka Oblast’, Ukraine
orcid

Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

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

Representative Publications

View All Publications
1 July 2025 in Knowledge Economy and Lifelong Learning
Verified
Journal Article
Open access
PDF file
AZAndrii Zakharchenko
Andrii Zakharchenko
Kushnerov OleksandrKushnerov Oleksandr
Kushnerov Oleksandr
ZKZuzana Kubaščikova
Zuzana Kubaščikova
KBKristina Babenko
Kristina Babenko
This article presents the development and implementation of a web-based system for forecasting employee turnover using machine learning techniques. Staff turnover is a critical issue for modern organisations, influenced by a range of internal factors such as job satisfaction, tenure, and promotion history, as well as external elements like labour market dynamics and economic trends. The proposed system leverages a neural network trained on a Kaggle dataset containing employment data from a large American company. Key variables include satisfaction scores, salary levels, promotion status, and work hours. Data preprocessing involved transforming categorical variables and normalising inputs to prepare for binary classification. A neural network with three hidden layers was constructed and optimised using the Adam algorithm and binary cross-entropy loss function. The model achieved a forecasting accuracy of 87% on the test set, indicating strong predictive capability. A web application was developed to enable real-time analysis, allowing HR professionals to input employee data and receive immediate risk assessments. This tool supports strategic decision-making by identifying high-risk individuals and enabling timely interventions to reduce turnover. The system is adaptable, integrates easily with existing HR platforms, and improves efficiency by automating data analysis. Challenges related to data privacy and quality were addressed through careful design and compliance with data protection standards. Implementing such an intelligent forecasting system offers economic and organisational benefits, including reduced recruitment costs, enhanced workforce stability, and improved HR decision-making. The solution demonstrates significant potential for adoption in diverse organisational contexts seeking to optimise their human capital management strategies.

21 May 2026
Verified
Conference Paper
Closed access
Kushnerov OleksandrKushnerov Oleksandr
Kushnerov Oleksandr
VSVladyslav Sokol
Vladyslav Sokol
OAOlena Akhiiezer
Olena Akhiiezer
ACAndrii Cherniavskyi
Andrii Cherniavskyi
MNMaksym Nadtochyi
Maksym Nadtochyi
DBDmytro Balagura
Dmytro Balagura
The rapid transition from passive large language models to autonomous agent-based coding systems, including frameworks like Claude Code and OpenClaw, has significantly accelerated software development cycles while simultaneously complicating the enterprise attack surface. Conventional security paradigms relying on perimeter defenses and static role-based access control (RBAC) are inherently designed for deterministic entities, leaving them critically underequipped to mitigate threats posed by non-deterministic algorithmic agents. These autonomous systems frequently generate unpredictable, multi-step sequences of system calls upon ingesting poisoned external environment data or hidden prompt injections. This article proposes a novel Zero Trust architectural framework, complete with rigorous mathematical formalization, specifically engineered for secure multi-agent ecosystems. The architectural core features an intelligent Semantic Gateway deeply integrated with operating system-level coordination modules to dynamically provision ephemeral sandboxing environments. To achieve continuous authorization for asynchronous transactions, the framework utilizes a mathematical structure based on Heterogeneous Graph Neural Networks (HGNN). This algorithm continuously computes an agent's dynamic trust tensor by quantifying the topological divergence between its predicted and actual traversal paths within the deterministic space of a corporate repository graph. To empirically validate the analytical efficacy of this advanced architecture, the research team executed a large-scale simulation processing 100,000 asynchronous transactions. The proposed framework demonstrated exceptional predictive power, achieving a true positive rate (TPR) exceeding 98.7 % in the detection and preventive isolation of critical remote code execution vulnerabilities, including sophisticated CVE-2025-61260 exploits. Crucially, this robust security posture was maintained with a minimal false positive rate (FPR) of 1.1 % and an average processing overhead of just 45 milliseconds. Ultimately, this paradigm modernizes the NIST SP 800-207A standard by replacing static binary access controls with a continuous Markov decision process, transforming stochastic AI unpredictability into a quantifiable engineering asset for critical infrastructures.

27 June 2025
Verified
Conference Paper
Closed access
Kushnerov OleksandrKushnerov Oleksandr
Kushnerov Oleksandr
YMYevhen Melenti
Yevhen Melenti
SYSerhii Yevseiev
Serhii Yevseiev
VSVladyslav Sokol
Vladyslav Sokol
SMStanislav Milevskyi
Stanislav Milevskyi
SDSergii Dunaiev
Sergii Dunaiev
The multilayer perceptron (MLP) is an effective tool for classifying network anomalies. This paper examines the methodology for applying MLP to this task, covering data preprocessing, neural network architecture selection and configuration, and training and optimization using modern techniques. The KDD Cup 99 dataset is used as an example. Despite the emergence of more sophisticated deep learning architectures, MLP remains relevant due to its implementation simplicity, computational efficiency, and potential for high accuracy with proper data preparation and tuning. The paper also discusses the challenges of integrating MLP and other machine learning methods into operational intrusion detection systems, such as processing large-scale data streams, handling class imbalance, adapting to evolving threats, and interpreting model decisions. Future research directions include improving preprocessing, investigating hybrid models, enhancing robustness, and exploring explainable AI and federated learning paradigms for MLP application in distributed networks.

25 August 2025 in Економіка та суспільство
Verified
Journal Article
Open access
PDF file
Kushnerov OleksandrKushnerov Oleksandr
Kushnerov Oleksandr
У статті досліджуються глибокі структурні зміни ринку праці України у період 2017–2024 років під впливом пандемії COVID-19 та впливу повномасштабної війни. Мета полягає у системному аналізі динаміки вакансій, рівня безробіття й дисбалансу між попитом і пропозицією робочої сили з використанням офіційних статистичних даних Державної служби статистики України та Порталу відкритих даних. Автори досліджують секторальні трансформації, визначають області з найбільшим кадровим дефіцитом та надлишком, зокрема у технічних, аграрних і оборонних галузях. Значна увага приділяється наслідкам кадрових диспропорцій для інноваційного потенціалу країни, що стримують темпи впровадження новітніх технологій. У дослідженні враховано вплив соціальних та демографічних факторів, міграційних процесів і мобілізації, які спричинили істотні зміни в структурі зайнятості.

4 February 2026 in Information
Verified
Journal Article
Open access
PDF file
Kushnerov OleksandrKushnerov Oleksandr
Kushnerov Oleksandr
Ruslan ShevchukRuslan Shevchuk
Ruslan Shevchuk
Serhii YevseievSerhii Yevseiev
Serhii Yevseiev
Mikolaj KarpinskiMikolaj Karpinski
Mikolaj Karpinski
The rapid adoption of large language models (LLMs) in corporate and governmental systems has raised critical security concerns, particularly prompt injection attacks exploiting LLMs’ inability to differentiate control instructions from untrusted user inputs. This study systematically benchmarks neural network architectures for malicious prompt detection, emphasizing robustness against character-level adversarial perturbations—an aspect that remains comparatively underemphasized in the specific context of prompt-injection detection despite its established significance in general adversarial NLP. Using the Malicious Prompt Detection Dataset (MPDD) containing 39,234 labeled instances, eight architectures—Dense DNN, CNN, BiLSTM, BiGRU, Transformer, ResNet, and character-level variants of CNN and BiLSTM—were evaluated based on standard performance metrics (accuracy, F1-score, and AUC-ROC), adversarial robustness coefficients against spacing and homoglyph perturbations, and inference latency. Results indicate that the word-level 3_Word_BiLSTM achieved the highest performance on clean samples (accuracy = 0.9681, F1 = 0.9681), whereas the Transformer exhibited lower accuracy (0.9190) and significant vulnerability to spacing attacks (adversarial robustness ρspacing=0.61). Conversely, the Character-level BiLSTM demonstrated superior resilience (ρspacing=1.0, ρhomoglyph=0.98), maintaining high accuracy (0.9599) and generalization on external datasets with only 2–4% performance decay. These findings highlight that character-level representations provide intrinsic robustness against obfuscation attacks, suggesting Char_BiLSTM as a reliable component in defense-in-depth strategies for LLM-integrated systems.

7 November 2024
Verified
Conference Paper
Closed access
Kushnerov OleksandrKushnerov Oleksandr
Kushnerov Oleksandr
PMPierre Murr
Pierre Murr
SHSerhii Herasymov
Serhii Herasymov
SMStanislav Milevskyi
Stanislav Milevskyi
MMMarharyta Melnyk
Marharyta Melnyk
SGSergii Golovashych
Sergii Golovashych
Monitoring and analyzing network traffic are critically important tasks in the context of increasing cyber threats. Traditional data processing methods often struggle to handle the large volumes and complexity of modern network traffic, necessitating the use of advanced technologies such as neural networks and machine learning. This article demonstrates the application of neural networks for monitoring and analyzing network traffic, highlighting their ability to learn from large datasets to detect complex patterns and anomalies. Using TensorFlow, we developed a neural network model based on an autoencoder that effectively identifies and classifies anomalies in network traffic. The model's effectiveness was validated using a synthetic dataset from Kaggle, showing high accuracy on both training and testing data. Our approach significantly improves the efficiency and accuracy of threat detection compared to traditional methods, making it a valuable tool for enhancing cybersecurity.