Towards transferable adversarial attacks on vision transformers for image classification

Xu Guo, Peng Chen, Zhihui Lu, Hongfeng Chai, Xin Du, Xudong Wu

Abstract: Highlights•Investigation of the vulnerability of Vision Transformer (ViT) to transfer attacks.•Introducing a novel Forward-Backward Transfer Adversarial Attack (FBTA) framework.•Outperformance of existing state-of-the-art methods on various models using the ImageNet validation dataset.•Studying the success rate of transfer attacks against quantized models.•Contribution to providing significant implications for the development of secure Artificial Intelligence (AI) systems in fintech regulation.