Adaptive Equilibrium: Dynamic Weighting Framework for Generalized Interruption of DeepFake Models
For researchers in adversarial attacks and deepfake defense, this work tackles the bottleneck of interruption imbalance in universal perturbation generation, offering a more balanced approach.
The paper addresses the problem of imbalance in generalized deepfake disruption, where static gradient normalization biases optimization towards susceptible models. The proposed Adaptive Equilibrium Framework (AEF) uses dynamic weighting based on real-time loss feedback to achieve uniform effectiveness, maintaining consistent interruption success across diverse architectures.
The advancement of generalized deepfake disruption is constrained by the interruption imbalance, a fundamental bottleneck inherent to the generation of universal perturbations. We reveal that conventional static gradient normalization fundamentally struggles to resolve architectural conflicts, causing the optimization to bias towards susceptible models while neglecting resistant ones. We argue that achieving high and uniform effectiveness requires resolving this imbalance by reaching an adaptive equilibrium. We propose the Adaptive Equilibrium Framework (AEF), which employs a dynamic weighting mechanism that utilizes real-time loss feedback to adaptively assign greater interruption weights to the most resistant models. This approach shifts the optimization from an average-case problem to finding a dynamic balance, driving the perturbation to a uniformly effective equilibrium state. Comprehensive experiments validate that AEF achieves a more balanced interruption performance, maintaining a consistent interruption success rate across the evaluated diverse architectures.