Improving Adversarial Robustness of Masked Autoencoders via Test-time Frequency-domain Prompting
This work addresses adversarial robustness for vision transformers using BERT pretraining, offering an incremental improvement for domain-specific applications in computer vision.
The paper tackles the poor adversarial robustness of Masked Autoencoders (MAE) compared to other BERT pretraining methods by identifying that pixel-level reconstruction targets degrade robustness by focusing on medium-/high-frequency components. It proposes a test-time frequency-domain prompting method that uses dataset-extracted prompts to occupy these frequencies, boosting MAE's adversarial robustness while maintaining clean performance on ImageNet-1k classification.
In this paper, we investigate the adversarial robustness of vision transformers that are equipped with BERT pretraining (e.g., BEiT, MAE). A surprising observation is that MAE has significantly worse adversarial robustness than other BERT pretraining methods. This observation drives us to rethink the basic differences between these BERT pretraining methods and how these differences affect the robustness against adversarial perturbations. Our empirical analysis reveals that the adversarial robustness of BERT pretraining is highly related to the reconstruction target, i.e., predicting the raw pixels of masked image patches will degrade more adversarial robustness of the model than predicting the semantic context, since it guides the model to concentrate more on medium-/high-frequency components of images. Based on our analysis, we provide a simple yet effective way to boost the adversarial robustness of MAE. The basic idea is using the dataset-extracted domain knowledge to occupy the medium-/high-frequency of images, thus narrowing the optimization space of adversarial perturbations. Specifically, we group the distribution of pretraining data and optimize a set of cluster-specific visual prompts on frequency domain. These prompts are incorporated with input images through prototype-based prompt selection during test period. Extensive evaluation shows that our method clearly boost MAE's adversarial robustness while maintaining its clean performance on ImageNet-1k classification. Our code is available at: https://github.com/shikiw/RobustMAE.