MIG-COW Adversarial Attack Achieves 99.96% White-Box Success on Deepfake Detectors
October 23, 2026
MIG-COW uses momentum-integrated gradients and consensus-orthogonal decomposition to attack both CNN and ViT-based deepfake detectors. While achieving near-perfect white-box success on AADD-2025 benchmarks, black-box transferability remains low at 7.16%.
HOW THIS AFFECTS YOU
●
researcherYou should evaluate your deepfake detectors against gradient decomposition attacks to test robustness.
●
policyThe high white-box success rate highlights the persistent fragility of current automated media authentication systems.