Reproduce “LOCALINTEL: Generating Organizational Threat Intelligence from Global and Local Cyber Knowledge” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Shaswata Mitra, Subash Neupane, Trisha Chakraborty, Sudip Mittal, Aritran Piplai, Manas Gaur, Shahram Rahimi
LocalIntel introduces a framework that automates the contextualization of global cyber threat intelligence by integrating it with an organization's private local knowledge using Retrieval Augmented Generation (RAG) and Large Language Models (LLMs). This system generates real-time, organization-specific threat intelligence reports, including tailored implications and mitigation strategies, achieving up to 93% accuracy and 78% correctness in evaluations.
Key insights
- Combining Retrieval Augmented Generation with LLMs effectively bridges the gap between generic global threat information and specific organizational contexts, leading to actionable intelligence.
- An iterative retrieval process that considers both global and local knowledge is crucial for accurately identifying threats relevant to an organization and generating precise mitigation strategies.
- Automating the contextualization of cyber threat intelligence significantly reduces the manual workload for SoC analysts, enabling faster and more accurate responses to evolving threats.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.