Backdoor Attacks to Pre-trained Unified Foundation Models. (arXiv:2302.09360v1 [cs.CR])

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 …

attack attacks backdoor backdoor attacks concept down foundation foundation models making modal proof-of-concept research security serious single support threat