Membership Inference on Text-to-Image Diffusion Models via Conditional Likelihood Discrepancy

TL;DR

This paper introduces CLiD, a new method for membership inference on text-to-image diffusion models that leverages conditional likelihood discrepancy to detect data privacy leakage effectively.

Contribution

The paper identifies a conditional overfitting phenomenon in text-to-image diffusion models and proposes CLiD, an analytical indicator for membership inference that outperforms previous methods.

Findings

CLiD significantly outperforms previous methods in various datasets.

The method is robust against overfitting mitigation strategies.

It effectively detects unauthorized data usage in diffusion models.

Abstract

Text-to-image diffusion models have achieved tremendous success in the field of controllable image generation, while also coming along with issues of privacy leakage and data copyrights. Membership inference arises in these contexts as a potential auditing method for detecting unauthorized data usage. While some efforts have been made on diffusion models, they are not applicable to text-to-image diffusion models due to the high computation overhead and enhanced generalization capabilities. In this paper, we first identify a conditional overfitting phenomenon in text-to-image diffusion models, indicating that these models tend to overfit the conditional distribution of images given the corresponding text rather than the marginal distribution of images only. Based on this observation, we derive an analytical indicator, namely Conditional Likelihood Discrepancy (CLiD), to perform membership inference, which reduces the stochasticity in estimating memorization of individual samples. Experimental results demonstrate that our method significantly outperforms previous methods across various data distributions and dataset scales. Additionally, our method shows superior resistance to overfitting mitigation strategies, such as early stopping and data augmentation.