CTI-HAL is a novel human-annotated dataset designed to advance automated extraction of cyber threat intelligence from unstructured text sources.

• Created to improve AI-based analysis of Advanced Persistent Threats (APTs)
• Includes 400 documents with 4,000+ annotations mapped to the MITRE ATT&CK framework
• Achieves higher annotation quality and more consistent coverage than existing datasets
• Enables more accurate threat detection and classification systems

This research addresses a critical security challenge: automating the extraction of actionable intelligence from vast amounts of unstructured threat data, helping organizations better prepare for and respond to sophisticated cyber attacks.

CTI-HAL: A Human-Annotated Dataset for Cyber Threat Intelligence Analysis