MOAT: Defending ViTs from Efficiency Attacks
Abstract
To adopt the Vision Transformers (ViTs) in resource-constrained environment, token pruning [1], [2] is widely used to reduce computational cost without impacting accuracy. However, adversaries have developed targeted attacks against said token pruning techniques to undermine such attempts to make ViTs efficient [3], [4]. In this paper, we propose MOAT (MOdel Agnostic randomized Transformations), a modelagnostic pre-processing defense pipeline that applies a combination of input transformations to protect efficient ViT implementations against adversarial efficiency attacks. MOAT operates directly on the input without requiring modifications to the model architecture or token pruning mechanism. Experimental results demonstrate that, across all evaluated ViT models, MOAT limits GFLOPs degradation under adversarial attacks to within 3.4% of the original unattacked model.