AMDS: Attack-Aware Multi-Stage Defense System for Network Intrusion Detection with Two-Stage Adaptive Weight Learning

Researchers developed AMDS, an attack-aware multi-stage defense system for network intrusion detection that uses adaptive weight learning to counter adversarial attacks. The system achieved 94.2% AUC and improved classification accuracy by 4.5 percentage points over existing adversarially trained ensembles by learning attack-specific detection strategies.

• →AMDS uses a weighted combination of ensemble disagreement, predictive uncertainty, and distributional anomaly signals for attack detection.
• →The system demonstrated 94.2% area under ROC curve and maintained 94.4% accuracy under adaptive white-box attacks with only 4.2% attack success rate.
• →Empirical analysis across seven adversarial attack types revealed distinct detection signatures enabling two-stage adaptive detection.
• →The defense framework improved F1-score by 9.0 points compared to adversarially trained ensembles on benchmark datasets.
• →Cross-dataset validation showed defense effectiveness depends on baseline classifier competence and may vary with feature dimensionality.