Understanding the Vulnerability of CLIP to Image Compression

Cangxiong Chen, Vinay P. Namboodiri, Julian Padget

Keywords: CLIP, vulnerability, zero-shot learning, image compression, attribution method, Integrated Gradients

TL;DR: We apply an attribution method called Integrated Gradients to understand the vulnerability of CLIP under image compression

Abstract: CLIP is a widely used foundational vision-language model that is used for zero-shot image recognition and other image-text alignment tasks. We demonstrate that CLIP is vulnerable to change in image quality under compression. This surprising result is further analysed using an attribution method-Integrated Gradients. Using this attribution method, we are able to better understand both quantitatively and qualitatively exactly the nature in which the compression affects the zero-shot recognition accuracy of this model. We evaluate this extensively on CIFAR-10 and STL-10. Our work provides the basis to understand this vulnerability of CLIP and can help us develop more effective methods to improve the robustness of CLIP and other vision-language models.

Submission Number: 91