This research introduces a novel approach to automate vulnerability evaluation in medical devices using large language models, addressing the growing cybersecurity challenge of rapidly assessing new vulnerabilities.

• Leverages historical vulnerability assessments to train LLMs for faster evaluation
• Integrates domain-specific ontologies to enhance accuracy and relevance
• Demonstrates effective automation for medical device security within a manufacturer's portfolio
• Projects significant time and resource savings in vulnerability management

This advancement is critical for security professionals facing the increasing volume of vulnerabilities published monthly in the National Vulnerability Database, with a projected 25% increase in 2024.

CVE-LLM: Ontology-Assisted Automatic Vulnerability Evaluation Using Large Language Models