This research introduces a novel approach using Large Language Models to automate the evaluation of security vulnerabilities in medical devices, addressing the growing challenge of rapidly identifying critical cybersecurity risks.

• Leverages historical vulnerability assessments to train LLMs for accurate security evaluation
• Integrates domain-specific ontology knowledge to enhance assessment quality
• Demonstrates practical application within a medical device manufacturer's cybersecurity workflow
• Provides an efficient solution for managing the projected 25% increase in new vulnerabilities by 2024

This innovation helps security teams prioritize critical vulnerabilities faster, reducing potential breach impacts and compliance risks in highly regulated medical environments.

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