Serhii Buchyk
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Serhii Buchyk has a total of 37 co-authors. Here are the top co-authors by number of shared publications:
- Ruslana Ziubina
Taras Shevchenko National University of Kyiv
5 shared publications Oleksandr Yudin
Taras Shevchenko National University of Kyiv Volodymyrska str., 60, Kyiv, Ukraine, 01601
4 shared publications- Serhii Toliupa
Taras Shevchenko National University of Kyiv
3 shared publications - Vitalii Piatyhor
Taras Shevchenko National University of Kyiv
3 shared publications - Anatolii Shevchenko
Taras Shevchenko National University of Kyiv
2 shared publications - Andrii Kulko
Taras Shevchenko National University of Kyiv
2 shared publications
Latest Publications
- Open Access
- 20 November 2026
- Published by European Scientific Platform (Publications)
- p. 187-189
ARCHITECTURE OF A HYBRID SYSTEM FOR AUTOMATIC MALWARE UNPACKING
- 20 November 2026
- Published by European Scientific Platform (Publications)
- p. 187-189
Open AccessARCHITECTURE OF A HYBRID SYSTEM FOR AUTOMATIC MALWARE UNPACKING
- Open Access
- 25 June 2026
- Published by Borys Grinchenko Kyiv University in Cybersecurity: Education, Science, Technique
- Vol. 1 (33) , 728-744
METHODOLOGICAL PRINCIPLES FOR IDENTIFYING BOT ACCOUNTS ON SOCIAL MEDIA
This article examines the methodological principles for identifying automated social media accounts. The relevance of the study is determined by the growing role of bots in the dissemination of misinformation, the artificial amplification of content, the manipulation of public opinion, and coordinated information campaigns. The emergence of generative artificial intelligence further complicates detection because automated accounts can produce natural language, vary their activity patterns, and imitate legitimate users more effectively. The reviewed approaches are systematized according to the primary source of features used for classification: user profile attributes, behavioral patterns, textual content, graph structures, and hybrid combinations of heterogeneous data. The methods are compared with respect to model type, datasets, reported evaluation metrics, data acquisition complexity, limitations, and explainability of classification decisions. Profile-based methods are shown to be scalable and relatively interpretable because they rely on accessible account metadata, but they may be insufficient for detecting sophisticated bots that maintain credible profiles. Behavioral approaches can reveal abnormal posting rhythms, repetitive activity, and coordination, although they require a sufficiently long activity history and are often applied to groups of accounts rather than individual users. Text-based methods can achieve strong classification results by analyzing message content through recurrent neural networks or transformer representations; however, they are sensitive to language, topic, dataset composition, and the increasing naturalness of AI-generated text. Graph-based characteristics are valuable for identifying coordinated amplification and bot networks: the reviewed evidence indicates differences between bot and human interaction structures, including denser bot ego-networks and a greater proportion of bot-to-bot links. At the same time, graph-based analysis is constrained by the cost, incompleteness, and limited reproducibility of large-scale interaction data acquisition. The analysis of hybrid approaches shows that combining profile, behavioral, textual, and, where available, graph features provides a more comprehensive representation of an account and can reduce dependence on a single weak modality. Nevertheless, complex multi-source models usually require more computational resources and have lower inherent explainability. The review identifies key unresolved problems: bot adaptability, heterogeneous and aging datasets, class imbalance, limited access to platform data, non-comparability of results obtained on different benchmarks, and insufficient interpretation of complex models. The results support the development of explainable hybrid detection models that combine accessible profile, behavioral, and textual features without critical dependence on a complete social graph.
- 25 June 2026
- Published by Borys Grinchenko Kyiv University in Cybersecurity: Education, Science, Technique
- Vol. 1 (33) , 728-744
Open AccessMETHODOLOGICAL PRINCIPLES FOR IDENTIFYING BOT ACCOUNTS ON SOCIAL MEDIA
This article examines the methodological principles for identifying automated social media accounts. The relevance of the study is determined by the growing role of bots in the dissemination of misinformation, the artificial amplification of content, the manipulation of public opinion, and coordinated information campaigns. The emergence of generative artificial intelligence further complicates detection because automated accounts can produce natural language, vary their activity patterns, and imitate legitimate users more effectively. The reviewed approaches are systematized according to the primary source of features used for classification: user profile attributes, behavioral patterns, textual content, graph structures, and hybrid combinations of heterogeneous data. The methods are compared with respect to model type, datasets, reported evaluation metrics, data acquisition complexity, limitations, and explainability of classification decisions. Profile-based methods are shown to be scalable and relatively interpretable because they rely on accessible account metadata, but they may be insufficient for detecting sophisticated bots that maintain credible profiles. Behavioral approaches can reveal abnormal posting rhythms, repetitive activity, and coordination, although they require a sufficiently long activity history and are often applied to groups of accounts rather than individual users. Text-based methods can achieve strong classification results by analyzing message content through recurrent neural networks or transformer representations; however, they are sensitive to language, topic, dataset composition, and the increasing naturalness of AI-generated text. Graph-based characteristics are valuable for identifying coordinated amplification and bot networks: the reviewed evidence indicates differences between bot and human interaction structures, including denser bot ego-networks and a greater proportion of bot-to-bot links. At the same time, graph-based analysis is constrained by the cost, incompleteness, and limited reproducibility of large-scale interaction data acquisition. The analysis of hybrid approaches shows that combining profile, behavioral, textual, and, where available, graph features provides a more comprehensive representation of an account and can reduce dependence on a single weak modality. Nevertheless, complex multi-source models usually require more computational resources and have lower inherent explainability. The review identifies key unresolved problems: bot adaptability, heterogeneous and aging datasets, class imbalance, limited access to platform data, non-comparability of results obtained on different benchmarks, and insufficient interpretation of complex models. The results support the development of explainable hybrid detection models that combine accessible profile, behavioral, and textual features without critical dependence on a complete social graph.
- Open Access
- 8 May 2026
- Published by European Scientific Platform (Publications)
- No. May 8, 202,p. 106-109
A COMPARISON OF SAST AND DAST IN DETECTING CRITICAL WEB VULNERABILITIES IN A VIRTUAL ENVIRONMENT
With the growth of web technologies and the widespread adoption of web applications in mission-critical industries, the number and complexity of cyberattacks are steadily increasing. According to the OWASP Top 10 [1], the vast majority of successful attacks on web systems exploit well-known classes of vulnerabilities.
- 8 May 2026
- Published by European Scientific Platform (Publications)
- No. May 8, 202,p. 106-109
Open AccessA COMPARISON OF SAST AND DAST IN DETECTING CRITICAL WEB VULNERABILITIES IN A VIRTUAL ENVIRONMENT
With the growth of web technologies and the widespread adoption of web applications in mission-critical industries, the number and complexity of cyberattacks are steadily increasing. According to the OWASP Top 10 [1], the vast majority of successful attacks on web systems exploit well-known classes of vulnerabilities.
- Open Access
- 26 March 2026
- Published by Borys Grinchenko Kyiv University in Cybersecurity: Education, Science, Technique
- p. 924
SOCIAL MEDIA BOT ACCOUNT DETECTION MODEL BASED ON MULTI-VIEW METHOD WITH ATTENTION MECHANISM
The article proposes a multi-view model for detecting bot accounts, which is based on the use of the ensemble Gradient Boosting method and the attention mechanism for integrating the results of the analysis of individual representations. The proposed approach involves considering each type of data as a separate representation, for which its own feature vector is formed, and a local model is trained. The behavioral representation considers the temporal characteristics of user activity, the attributive representation - the properties of the profile, the content representation - the textual features of posts using semantic models, and the visual representation - the features of images, including metadata analysis and OCR. For each representation, a local estimate of the probability of automated activity and a data quality indicator are calculated. The integration of results is performed using an attention mechanism that determines the weight of each representation depending on its informativeness and reliability. This allows us to adaptively take into account incompleteness or heterogeneity of data and increases the model's resilience to the absence of individual types of information. The final score is formed based on a weighted combination of representations using the Gradient Boosting metamodel. A feature of the proposed approach is the ability to interpret the results by assessing the contribution of each representation and determining the level of confidence of the model. This ensures transparency of the decisions made and allows the model to be used in practical social network analysis systems. The proposed approach extends existing bot detection methods by combining multi-species learning, adaptive aggregation, and data quality consideration, which increases the efficiency of the developed system.
- 26 March 2026
- Published by Borys Grinchenko Kyiv University in Cybersecurity: Education, Science, Technique
- p. 924
Open AccessSOCIAL MEDIA BOT ACCOUNT DETECTION MODEL BASED ON MULTI-VIEW METHOD WITH ATTENTION MECHANISM
The article proposes a multi-view model for detecting bot accounts, which is based on the use of the ensemble Gradient Boosting method and the attention mechanism for integrating the results of the analysis of individual representations. The proposed approach involves considering each type of data as a separate representation, for which its own feature vector is formed, and a local model is trained. The behavioral representation considers the temporal characteristics of user activity, the attributive representation - the properties of the profile, the content representation - the textual features of posts using semantic models, and the visual representation - the features of images, including metadata analysis and OCR. For each representation, a local estimate of the probability of automated activity and a data quality indicator are calculated. The integration of results is performed using an attention mechanism that determines the weight of each representation depending on its informativeness and reliability. This allows us to adaptively take into account incompleteness or heterogeneity of data and increases the model's resilience to the absence of individual types of information. The final score is formed based on a weighted combination of representations using the Gradient Boosting metamodel. A feature of the proposed approach is the ability to interpret the results by assessing the contribution of each representation and determining the level of confidence of the model. This ensures transparency of the decisions made and allows the model to be used in practical social network analysis systems. The proposed approach extends existing bot detection methods by combining multi-species learning, adaptive aggregation, and data quality consideration, which increases the efficiency of the developed system.
- Open Access
- 26 September 2025
- Published by Borys Grinchenko Kyiv University in Cybersecurity: Education, Science, Technique
- Vol. 1 (29) , 687-695
COMPREHENSIVE MODEL FOR SELECTING ANTIVIRUS SOFTWARE FOR CRITICAL INFRASTRUCTURE FACILITIES
The rapid growth of cyber attacks on critical infrastructure against the backdrop of rapid development of information technology highlights the urgent need for effective protection mechanisms. This study examines the problem of choosing the best antivirus software, which is complicated by the increasing complexity of decision-making and the limitations of current approaches. The article presents a detailed, comprehensive mathematical model designed to automate the process of selecting antivirus software. This model uses multi-criteria analysis and expert assessments to ensure flexibility and objectivity. It evaluates technical characteristics such as detection probability, false positive rate, performance, system load, response time, and database update frequency, as well as operational risks and the Common Vulnerability Scoring System (CVSS). The methodology uses geometric mean aggregation of expert opinions and checks their consistency to reduce subjectivity and increase reliability. Two methods of incorporating the risk factor into the overall assessment were tested, allowing adaptation to different strategic goals. Experimental results confirmed the effectiveness of the model in determining optimal antivirus solutions for mission-critical information systems. The model effectively processes expert data and creates organised information for decision-making. Further research could expand the model's knowledge base, improve decision-making algorithms for new cyber threats, and adapt it for wider application in information security.
- 26 September 2025
- Published by Borys Grinchenko Kyiv University in Cybersecurity: Education, Science, Technique
- Vol. 1 (29) , 687-695
Open AccessCOMPREHENSIVE MODEL FOR SELECTING ANTIVIRUS SOFTWARE FOR CRITICAL INFRASTRUCTURE FACILITIES
The rapid growth of cyber attacks on critical infrastructure against the backdrop of rapid development of information technology highlights the urgent need for effective protection mechanisms. This study examines the problem of choosing the best antivirus software, which is complicated by the increasing complexity of decision-making and the limitations of current approaches. The article presents a detailed, comprehensive mathematical model designed to automate the process of selecting antivirus software. This model uses multi-criteria analysis and expert assessments to ensure flexibility and objectivity. It evaluates technical characteristics such as detection probability, false positive rate, performance, system load, response time, and database update frequency, as well as operational risks and the Common Vulnerability Scoring System (CVSS). The methodology uses geometric mean aggregation of expert opinions and checks their consistency to reduce subjectivity and increase reliability. Two methods of incorporating the risk factor into the overall assessment were tested, allowing adaptation to different strategic goals. Experimental results confirmed the effectiveness of the model in determining optimal antivirus solutions for mission-critical information systems. The model effectively processes expert data and creates organised information for decision-making. Further research could expand the model's knowledge base, improve decision-making algorithms for new cyber threats, and adapt it for wider application in information security.
- Open Access
- 26 June 2025
- Published by Borys Grinchenko Kyiv University in Cybersecurity: Education, Science, Technique
- Vol. 4 (28) , 565-574
DETECTING PHISHING URLS BASED ON HEURISTIC RULES
The growing number of cyber threats, including phishing attacks, requires the development of effective phishing detection methods. Phishing is one of the most widespread and dangerous forms of cybercrime aimed at gaining access to confidential user information by manipulating their trust. The methodology of social engineering attacks is evolving rapidly, which creates challenges for cybersecurity professionals. This study discusses a rule-based heuristic approach to detecting phishing URLs that analyses various characteristics of web addresses, such as atypical link structure, suspicious characters, or incorrect word sequences in domain names, to identify potential threats. A mind map, as a visualisation tool, provides a structured and logically ordered approach to analysing phishing website detection methods. It clearly demonstrates the relationships between various rules aimed at identifying suspicious behaviour of web resources, and also allows you to clearly trace the classification of characteristics inherent in phishing attacks. This map summarises the key features used to analyse URLs, allowing for a structured presentation of the criteria for assessing a potential threat. Having a clear classification of heuristic detection rules, the accuracy of identifying phishing websites is significantly improved, protecting users from possible attacks. The study evaluates the effectiveness of individual heuristic rules and their combinations, which gives an idea of their applicability to automated phishing detection systems. For this purpose, we use key metrics that determine the quality of the algorithm and its ability to distinguish between phishing and legitimate websites, namely reliability, true positive rate, false positive rate, accuracy, and F-1 score. The results demonstrate the feasibility of heuristic detection as a simple, effective and understandable alternative to complex machine learning models.
- 26 June 2025
- Published by Borys Grinchenko Kyiv University in Cybersecurity: Education, Science, Technique
- Vol. 4 (28) , 565-574
Open AccessDETECTING PHISHING URLS BASED ON HEURISTIC RULES
The growing number of cyber threats, including phishing attacks, requires the development of effective phishing detection methods. Phishing is one of the most widespread and dangerous forms of cybercrime aimed at gaining access to confidential user information by manipulating their trust. The methodology of social engineering attacks is evolving rapidly, which creates challenges for cybersecurity professionals. This study discusses a rule-based heuristic approach to detecting phishing URLs that analyses various characteristics of web addresses, such as atypical link structure, suspicious characters, or incorrect word sequences in domain names, to identify potential threats. A mind map, as a visualisation tool, provides a structured and logically ordered approach to analysing phishing website detection methods. It clearly demonstrates the relationships between various rules aimed at identifying suspicious behaviour of web resources, and also allows you to clearly trace the classification of characteristics inherent in phishing attacks. This map summarises the key features used to analyse URLs, allowing for a structured presentation of the criteria for assessing a potential threat. Having a clear classification of heuristic detection rules, the accuracy of identifying phishing websites is significantly improved, protecting users from possible attacks. The study evaluates the effectiveness of individual heuristic rules and their combinations, which gives an idea of their applicability to automated phishing detection systems. For this purpose, we use key metrics that determine the quality of the algorithm and its ability to distinguish between phishing and legitimate websites, namely reliability, true positive rate, false positive rate, accuracy, and F-1 score. The results demonstrate the feasibility of heuristic detection as a simple, effective and understandable alternative to complex machine learning models.
- Open Access
- 1 January 2025
- Published by Taras Shevchenko National University of Kyiv in Information systems and technologies security
- Vol. 1 (1 (9)) , 11-17
Social media automated account (bot) detection system architecture
Background. Social networks are one of the main sources of information. However, as trust in them grows, so does the use of automated accounts to spread disinformation and influence public opinion, which is an effective method of manipulation. Automated account detection systems are used to detect such accounts. The purpose of the article is to analyze the architecture of automated account (bot) detection systems in social networks and to identify the main directions for improving such architecture, taking into account the identified shortcomings. Methods. Methods of analysis, systematization, and generalization were used to identify shortcomings and areas for system improvement, and a modeling method was used to develop generalized and improved system architecture. Results. An improved architecture of a system for detecting automated accounts (bots) in the social network x.com is proposed, which uses a mixed model of data analysis using artificial intelligence and combines web scraping and API queries for data collection. Conclusions. The general architecture of automated account (bot) detection systems in social networks was analyzed and the main shortcomings and areas for improvement were identified, among which the following can be distinguished: restrictions on API requests from social network owners, unstructured and free style of posts in social networks, and constant changes in post generation algorithms by automated accounts. An improved architecture of the automated account (bot) detection system was proposed, which combines web scraping and API requests for data collection and allows parallel execution of social network data analysis. Keywords: сomputer system, data collection, data analysis, machine learning.
- 1 January 2025
- Published by Taras Shevchenko National University of Kyiv in Information systems and technologies security
- Vol. 1 (1 (9)) , 11-17
Open AccessSocial media automated account (bot) detection system architecture
Background. Social networks are one of the main sources of information. However, as trust in them grows, so does the use of automated accounts to spread disinformation and influence public opinion, which is an effective method of manipulation. Automated account detection systems are used to detect such accounts. The purpose of the article is to analyze the architecture of automated account (bot) detection systems in social networks and to identify the main directions for improving such architecture, taking into account the identified shortcomings. Methods. Methods of analysis, systematization, and generalization were used to identify shortcomings and areas for system improvement, and a modeling method was used to develop generalized and improved system architecture. Results. An improved architecture of a system for detecting automated accounts (bots) in the social network x.com is proposed, which uses a mixed model of data analysis using artificial intelligence and combines web scraping and API queries for data collection. Conclusions. The general architecture of automated account (bot) detection systems in social networks was analyzed and the main shortcomings and areas for improvement were identified, among which the following can be distinguished: restrictions on API requests from social network owners, unstructured and free style of posts in social networks, and constant changes in post generation algorithms by automated accounts. An improved architecture of the automated account (bot) detection system was proposed, which combines web scraping and API requests for data collection and allows parallel execution of social network data analysis. Keywords: сomputer system, data collection, data analysis, machine learning.
- 20 March 2024
- Published by Academy of Cognitive and Natural Sciences
Methodology of FPGA Implementation and Performance Evaluation of Polar Coding for 5G Communications
This article is devoted to the study of noise immunity of an infocommunication system with Polar Coding (P-C). The results of evaluating the performance of the Successive Cancellation (S-C) decoder are presented. The main contribution of this paper is the result of the compromise between high latency and Field Programmable Gate Array (FPGA) resources to implement a P-C decoder for 5G communications. The stages of channel polarization, the mathematical description of P-C, and the features of their description in binary channels are presented separately. This is achieved by the method of mathematical modeling of information and statistical characteristics of P-C for different code configurations. To check the correctness of the decisions made, a comparative description of the advantages and disadvantages of P-C decoding algorithms is given. The specifics of the FPGA implementation of the S-C algorithm have been studied. Based on the results of the experiment, the noise immunity of the P-C channel was assessed when changing the length of the code block, adding Cyclic Redundancy Check (CRC), and reversing bits at different code rates. It is expected that the results will be useful in optimizing the design process of real P-C circuits.
- 20 March 2024
- Published by Academy of Cognitive and Natural Sciences
Methodology of FPGA Implementation and Performance Evaluation of Polar Coding for 5G Communications
This article is devoted to the study of noise immunity of an infocommunication system with Polar Coding (P-C). The results of evaluating the performance of the Successive Cancellation (S-C) decoder are presented. The main contribution of this paper is the result of the compromise between high latency and Field Programmable Gate Array (FPGA) resources to implement a P-C decoder for 5G communications. The stages of channel polarization, the mathematical description of P-C, and the features of their description in binary channels are presented separately. This is achieved by the method of mathematical modeling of information and statistical characteristics of P-C for different code configurations. To check the correctness of the decisions made, a comparative description of the advantages and disadvantages of P-C decoding algorithms is given. The specifics of the FPGA implementation of the S-C algorithm have been studied. Based on the results of the experiment, the noise immunity of the P-C channel was assessed when changing the length of the code block, adding Cyclic Redundancy Check (CRC), and reversing bits at different code rates. It is expected that the results will be useful in optimizing the design process of real P-C circuits.
- 1 January 2024
- Published by Taras Shevchenko National University of Kyiv in Information systems and technologies security
- Vol. 2 (2 (8)) , 17-25
JAVASCRIPT-BASED WEB EXPLOIT ANALYSIS MODEL
B a c k g r o u n d . The task of ensuring the security of web applications and servers remains important and relevant in the face of the ever-increasing number of attacks in cyberspace. The use of various open-source content management systems (e.g. WordPress, Joomla, Open Journal Systems, Drupal), which are quite popular for creating websites due to their ease of installation and use, unfortunately, require constant updating not only to improve the content but also to ensure the security of the system. In this article, the authors focus on the WordPress system, although this approach can be used for other systems as well. The article emphasises the importance of early detection of vulnerabilities to prevent potential cyber threats and their negative consequences. The article proposes a model and a script designed to speed up the detection of vulnerabilities in WordPress applications. Automation of the scanning process with a custom script allows you to quickly detect vulnerabilities, ensuring prompt fixes and updates. This approach not only strengthens security, but also helps preserve the reputation of websites and brands, which is critical in today's digital environment. M e t h o d s . The methods of analysing JavaScript-based web exploits were used, taking into account the general principles of their analysis and taking into account the methodologies for analysing web applications for vulnerabilities. R e s u l t s . An improved model of analysing a web application on CMS Wordpress based on a script that provides automated scanning of a web application by running the following utilities is presented: NMAP, Dirb, Nikto, SQLMap, WPScan and PwnXSS. All the results are recorded in a separate file for further study of all the found security issues of the web application. C o n c l u s i o n s . The developed model and script should help developers and testers speed up the process of identifying vulnerabilities in Wordpress, as they can run one script and get a voluminous and meaningful report with the identified vulnerabilities in a short time. This optimises vulnerability detection by automating the launch of scanners.
- 1 January 2024
- Published by Taras Shevchenko National University of Kyiv in Information systems and technologies security
- Vol. 2 (2 (8)) , 17-25
JAVASCRIPT-BASED WEB EXPLOIT ANALYSIS MODEL
B a c k g r o u n d . The task of ensuring the security of web applications and servers remains important and relevant in the face of the ever-increasing number of attacks in cyberspace. The use of various open-source content management systems (e.g. WordPress, Joomla, Open Journal Systems, Drupal), which are quite popular for creating websites due to their ease of installation and use, unfortunately, require constant updating not only to improve the content but also to ensure the security of the system. In this article, the authors focus on the WordPress system, although this approach can be used for other systems as well. The article emphasises the importance of early detection of vulnerabilities to prevent potential cyber threats and their negative consequences. The article proposes a model and a script designed to speed up the detection of vulnerabilities in WordPress applications. Automation of the scanning process with a custom script allows you to quickly detect vulnerabilities, ensuring prompt fixes and updates. This approach not only strengthens security, but also helps preserve the reputation of websites and brands, which is critical in today's digital environment. M e t h o d s . The methods of analysing JavaScript-based web exploits were used, taking into account the general principles of their analysis and taking into account the methodologies for analysing web applications for vulnerabilities. R e s u l t s . An improved model of analysing a web application on CMS Wordpress based on a script that provides automated scanning of a web application by running the following utilities is presented: NMAP, Dirb, Nikto, SQLMap, WPScan and PwnXSS. All the results are recorded in a separate file for further study of all the found security issues of the web application. C o n c l u s i o n s . The developed model and script should help developers and testers speed up the process of identifying vulnerabilities in Wordpress, as they can run one script and get a voluminous and meaningful report with the identified vulnerabilities in a short time. This optimises vulnerability detection by automating the launch of scanners.