Boosting Adversarial Transferability across Model Genus by Deformation-Constrained Warping
This addresses the problem of cross-model adversarial attacks for security researchers, offering a novel method to improve transferability across diverse model architectures, though it is incremental in the broader field of adversarial machine learning.
The paper tackles the limited transferability of adversarial examples across different model types (e.g., CNNs to Transformers) by proposing Deformation-Constrained Warping Attack (DeCoWA), which uses constrained elastic deformation to enhance transferability, achieving significant performance hindrance in tasks like image classification, video action recognition, and audio recognition.
Adversarial examples generated by a surrogate model typically exhibit limited transferability to unknown target systems. To address this problem, many transferability enhancement approaches (e.g., input transformation and model augmentation) have been proposed. However, they show poor performances in attacking systems having different model genera from the surrogate model. In this paper, we propose a novel and generic attacking strategy, called Deformation-Constrained Warping Attack (DeCoWA), that can be effectively applied to cross model genus attack. Specifically, DeCoWA firstly augments input examples via an elastic deformation, namely Deformation-Constrained Warping (DeCoW), to obtain rich local details of the augmented input. To avoid severe distortion of global semantics led by random deformation, DeCoW further constrains the strength and direction of the warping transformation by a novel adaptive control strategy. Extensive experiments demonstrate that the transferable examples crafted by our DeCoWA on CNN surrogates can significantly hinder the performance of Transformers (and vice versa) on various tasks, including image classification, video action recognition, and audio recognition. Code is made available at https://github.com/LinQinLiang/DeCoWA.