LARGE LANGUAGE MODELS FOR OPTIMIZING TASKS IN INFORMATION RETRIEVAL AND PROTECTION

Abstract

The rapid development of digital technologies, the global informatization of society, and the continuous growth in data volumes create new challenges in the field of information retrieval, processing, and protection. Information systems generate terabytes of data every day, including both valuable information and potential indicators of cyber threats, anomalies, and attacks. Traditional approaches to information analysis are increasingly proving insufficient due to their limited capability to process unstructured data and the need for significant human involvement.

The emergence of large language models (LLMs) has become one of the most significant achievements of modern artificial intelligence. Owing to their ability to analyze natural language, understand context, summarize vast amounts of information, and generate meaningful responses, LLMs provide new opportunities for automating information retrieval processes and enhancing information security. Recent studies demonstrate that large language models can be effectively applied to vulnerability detection, malware analysis, cyber incident investigation, security log processing, and the automation of Security Operations Center (SOC) activities.

This article investigates the potential of large language models for optimizing information retrieval, analysis, processing, and protection processes in the modern digital environment. The study examines the characteristics of large language models, the principles underlying their construction, and the mechanisms of their application within information and communication systems. The main areas of LLM utilization in information retrieval, automated text analysis, information threat detection, cyber incident monitoring, software vulnerability discovery, and decision-support processes in cybersecurity are analyzed. Particular attention is paid to the integration of large language models with Security Information and Event Management (SIEM) systems, intrusion detection systems, cyber threat intelligence technologies, and Retrieval-Augmented Generation (RAG) mechanisms. The advantages and limitations of employing LLMs in information retrieval and protection tasks are identified. Furthermore, the prospects for the further development of intelligent cyber defense systems based on large language models are substantiated.