The classifier judged this item not relevant to AI security, so it is left out of lists, feeds, the API and dataset releases. If that is wrong, report a correction on this page.
InfoResearchPeer-reviewed
ThreatMAMBA: Achieving High-Robustness Cyber Threat Attribution During the Evolution of Attacks
- Published
- Record updated
Summary
ThreatMAMBA is a framework for Cyber Threat Attribution that builds a heterogeneous information graph from Indicators of Compromise (IOCs), Tactics, Techniques, and Procedures (TTPs), and temporal relationships in Cyber Threat Intelligence (CTI). It uses a state space selection mechanism and contrastive learning to construct temporal knowledge graphs, improving the Macro F1 score by 6.94% to 12.05% across event progression stages from 20% to 100%. Its attacker behavior profiles reach a Jaccard similarity of 0.2991 with MITRE ATT&CK, outperforming Qi'anxin by 0.1836.