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
Causality-driven hybrid intrusion detection with entropy-aware feature learning and federated intelligence
- Published
- Record updated
Summary
A causality-driven hybrid intrusion detection framework combines five modules, including causal signature learning, entropy-guided feature compression and trust-weighted federated threat intelligence. Tested on CICIDS2017, UNSW-NB15 and IoT-23, it reported accuracies of 97.4%, 96.1% and 95.6% respectively, with 18 retained features (a 70.9% reduction) and 18.6 ms average inference time. Accuracy stayed above 91.5% under bounded gradient-based perturbations.