Safeguarding Data in Multimodal AI: A Differentially Private Approach to CLIP Training

Abstract

The surge in multimodal AI's success has sparked concerns over data privacy in vision-and-language tasks. While CLIP has revolutionized multimodal learning through joint training on images and text, its potential to unintentionally disclose sensitive information necessitates the integration of privacy-preserving mechanisms. We introduce a differentially private adaptation of the Contrastive Language-Image Pretraining (CLIP) model that effectively addresses privacy concerns while retaining accuracy. Our proposed method, \dpclip, is rigorously evaluated on benchmark datasets encompassing diverse vision-and-language tasks such as image classification and image captioning. We demonstrate that our approach retains performance on par with the standard non-private CLIP model. Furthermore, we analyze our proposed algorithm under linear representation settings. We derive the convergence rate of our algorithm and show a trade-off between utility and privacy when gradients are clipped per-\textit{batch} and the loss function does not satisfy smoothness conditions assumed in the literature for the analysis of DP-SGD.