AI RESEARCH
TIJERE: A Novel Threat Intelligence Joint Extraction Model Based on Analyst Expert Knowledge
arXiv CS.LG
•
ArXi:2605.02041v1 Announce Type: new The extraction of entities and relationships from threat intelligence reports into structured formats, such as cybersecurity knowledge graphs, is essential for automated threat analysis, detection, and mitigation. However, existing joint extraction methods struggle with feature confusion, language ambiguity, noise propagation, and overlapping relations, resulting in low accuracy and poor model performance.