LLM-assisted Cybersecurity Defense Framework (LLcydef)

Given the dynamic nature of IT security, it is crucial to stay ahead of the curve. With the rapid adoption of AI-powered tools, especially Large Language Models (LLMs), in the IT domain, it is becoming essential to leverage these technologies for cybersecurity. The "LLcydef" project proposes a cyber defense framework based on an LLM agent system. The aim is to detect and respond to security threats at an early stage. It is equipped with tools for security assessment, such as penetration tests. Simulated attacks are carried out to find vulnerabilities. The system learns attack patterns through synchronization and analysis of the Known Exploited Vulnerabilities Catalog (KEV). The "Expectation Maximization via Fourier Series" (EMoFS) method, which analyses complex data, is used to detect attacks. In the long term, the system is to be further developed so that it can also be used in areas such as machine learning and signal processing. In addition, a database for attack patterns is being set up and an existing language model for cyber security is being optimized. The aim is to proactively protect networks and identify vulnerabilities in real time.