The rise of pre-trained unified foundation models breaks down the barriers between different modalities and tasks, providing comprehensive support to users with unified architectures. However, the backdoor attack on pre-trained models poses a serious threat to their security. Previous research on backdoor attacks has been limited to uni-modal tasks or single tasks across modalities, making it inapplicable to unified foundation models. In this paper, we make proof-of-concept level research on the backdoor attack for pre-trained unified foundation models. Through preliminary experiments on NLP and CV classification tasks, we reveal the vulnerability of these models and suggest future research directions for enhancing the attack approach.
@article{arxiv.2302.09360,
title = {Backdoor Attacks to Pre-trained Unified Foundation Models},
author = {Zenghui Yuan and Yixin Liu and Kai Zhang and Pan Zhou and Lichao Sun},
journal= {arXiv preprint arXiv:2302.09360},
year = {2023}
}