Reproduce “ThreatKG: An AI-Powered System for Automated Open-Source Cyber Threat Intelligence Gathering and Management” 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

Peng Gao, Xiaoyuan Liu, Edward Choi, Sibo Ma, Xinyu Yang, Dawn Song

THREATKG is an AI-powered system developed by Virginia Tech and UC Berkeley that automatically extracts and manages cyber threat intelligence from unstructured open-source reports. The system constructs a unified threat knowledge graph, demonstrating an 85% F1-score in relation extraction and outperforming existing specialized extraction tools.

Key insights

  • A hierarchical threat knowledge ontology allows for a comprehensive and multi-dimensional representation of cyber threats, effectively linking low-level indicators with high-level attack concepts and their relationships.
  • Data programming, through noisy labeling functions and programmatic synthesis of annotations, provides an effective method to overcome the scarcity of labeled training data for advanced AI models in specialized domains like cybersecurity.
  • A modular and continuously updated architecture enables scalable and extensible management of evolving threat intelligence, allowing for seamless integration of new sources and adaptation to changing threat landscapes.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.