SleeperMark: Towards Robust Watermark against Fine-Tuning Text-to-image Diffusion Models

TL;DR

SleeperMark introduces a robust watermarking framework for text-to-image diffusion models that remains effective even after fine-tuning and various attacks, protecting intellectual property without compromising model performance.

Contribution

The paper presents SleeperMark, a novel method for embedding resilient watermarks into diffusion models that disentangle watermark information from semantic features, ensuring robustness against fine-tuning.

Findings

SleeperMark effectively preserves watermarks after fine-tuning.

The method demonstrates robustness against various attacks at image and model levels.

Minimal impact on the generative quality of the models.

Abstract

Recent advances in large-scale text-to-image (T2I) diffusion models have enabled a variety of downstream applications, including style customization, subject-driven personalization, and conditional generation. As T2I models require extensive data and computational resources for training, they constitute highly valued intellectual property (IP) for their legitimate owners, yet making them incentive targets for unauthorized fine-tuning by adversaries seeking to leverage these models for customized, usually profitable applications. Existing IP protection methods for diffusion models generally involve embedding watermark patterns and then verifying ownership through generated outputs examination, or inspecting the model's feature space. However, these techniques are inherently ineffective in practical scenarios when the watermarked model undergoes fine-tuning, and the feature space is inaccessible during verification ((i.e., black-box setting). The model is prone to forgetting the previously learned watermark knowledge when it adapts to a new task. To address this challenge, we propose SleeperMark, a novel framework designed to embed resilient watermarks into T2I diffusion models. SleeperMark explicitly guides the model to disentangle the watermark information from the semantic concepts it learns, allowing the model to retain the embedded watermark while continuing to be adapted to new downstream tasks. Our extensive experiments demonstrate the effectiveness of SleeperMark across various types of diffusion models, including latent diffusion models (e.g., Stable Diffusion) and pixel diffusion models (e.g., DeepFloyd-IF), showing robustness against downstream fine-tuning and various attacks at both the image and model levels, with minimal impact on the model's generative capability. The code is available at https://github.com/taco-group/SleeperMark.