ADAPT to Robustify Prompt Tuning Vision Transformers

arXiv:2403.13196v1 Announce Type: new
Abstract: The performance of deep models, including Vision Transformers, is known to be vulnerable to adversarial attacks. Many existing defenses against these attacks, such as adversarial training, rely on full-model fine-tuning to induce robustness in the models. These defenses require storing a copy of the entire model, that can have billions of parameters, for each task. At the same time, parameter-efficient prompt tuning is used to adapt large transformer-based models to downstream tasks without the need …

abstract adapt adversarial adversarial attacks adversarial training arxiv attacks copy cs.ai cs.cv cs.lg fine-tuning model fine-tuning performance prompt prompt tuning robustness stat.ml training transformers type vision vision transformers vulnerable