IntellBot: Retrieval Augmented LLM Chatbot for Cyber Threat Knowledge Delivery
Dincy R. Arikkat, M. Abhinav, Navya Binu, M. Parvathi, Navya Biju, K. S. Arunima, P. Vinod, Rafidha Rehiman K. A., Mauro Conti · 2024
In the evolving field of cybersecurity, AI-driven chatbots are valuable. Unlike rigid rule-based bots, Large Language Model (LLM) based chatbots can provide flexible, context-aware responses across various domains. This research introduces IntellBot, a sophisticated cybersecurity chatbot built using advanced LLMs, Langchain, and a Retrieval-Augmented Generation model. IntellBot aggregates data from diverse sources, creating a comprehensive knowledge base on vulnerabilities, recent cyber threats, and emerging trends. It offers intelligent responses to enhance threat intelligence, incident response, and overall security awareness. A two-stage evaluation yielded high accuracy, with a BERT score above 0.8 and cosine similarity from 0.8 to 1, while RAGAS assessments scores consistently over 0.77, confirming the system's effectiveness.