Olexander Puchkov,
Dmytro Lande,
Olexander Rybak
Integration of Technologies in Cybersecurity: Information Retrieval and Artificial Intelligence
// Selected Papers of the XXII International Scientific and Practical Conference
"Information Technologies and Security" (ITS 2022) CEUR Workshop Proceedings (ceur-ws.org).
- Vol-3887. - pp. 173-183. ISSN 1613-0073. [http://ceur-ws.org/Vol-3887/paper15.pdf]
Modern challenges in cybersecurity require new approaches to information retrieval and data analysis. The
growth of data volumes and the speed of their updates make traditional information processing methods
insufficiently effective. This paper proposes the integration of large language models (LLMs) into
information retrieval systems to enhance analytical capabilities and automate data processing tasks. The
main goal of the research is to translate the analytical component of the information retrieval system to
LLMs, significantly improving the accuracy, completeness, and relevance of information searches.
The system Cyber Aggregator, used for monitoring and analyzing social media content in the context of
cybersecurity, demonstrates the effectiveness of the proposed approach. The integration of LLMs into Cyber
Aggregator allows for the automation of semantic indexing processes, enhances the formulation and
modification of user queries, and provides more precise summarization and analysis of search results. This
includes creating analytical digests, identifying key events, constructing semantic maps, and conducting
semantic analysis.
The proposed methodology is based on leveraging the powerful capabilities of LLMs, such as understanding
complex relationships between concepts, analyzing context, and automatically forming conclusions. The
application of this technology in cybersecurity contributes to more effective threat monitoring, improved
situational awareness, and enhanced real-time threat response capabilities. The paper also presents a UML
diagram illustrating the key components of the system, along with a mathematical formalization of the main
processes related to the integration of LLMs into information retrieval systems.
The research findings indicate that the use of LLMs combined with information retrieval technologies opens
new opportunities for automating data analysis and ensuring cybersecurity. This makes the proposed
approach an important tool for cybersecurity professionals engaged in open-source intelligence (OSINT) and
other analytical tasks in today.s information environment.
Keywords
Cybersecurity, Information Retrieval, Large Language Models, Data Analysis Automation, Social Media
Monitoring, Semantic Analysis, Cyber Aggregator
|