AquaLoRA: Toward White-box Protection for Customized Stable Diffusion Models via Watermark LoRA

Weitao Feng, Wenbo Zhou, Jiyan He, Jie Zhang, Tianyi Wei, Guanlin Li, Tianwei Zhang, Weiming Zhang, Nenghai Yu
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:13423-13444, 2024.

Abstract

Diffusion models have achieved remarkable success in generating high-quality images. Recently, the open-source models represented by Stable Diffusion (SD) are thriving and are accessible for customization, giving rise to a vibrant community of creators and enthusiasts. However, the widespread availability of customized SD models has led to copyright concerns, like unauthorized model distribution and unconsented commercial use. To address it, recent works aim to let SD models output watermarked content for post-hoc forensics. Unfortunately, none of them can achieve the challenging white-box protection, wherein the malicious user can easily remove or replace the watermarking module to fail the subsequent verification. For this, we propose AquaLoRA as the first implementation under this scenario. Briefly, we merge watermark information into the U-Net of Stable Diffusion Models via a watermark LowRank Adaptation (LoRA) module in a two-stage manner. For watermark LoRA module, we devise a scaling matrix to achieve flexible message updates without retraining. To guarantee fidelity, we design Prior Preserving Fine-Tuning (PPFT) to ensure watermark learning with minimal impacts on model distribution, validated by proofs. Finally, we conduct extensive experiments and ablation studies to verify our design. Our code is available at github.com/Georgefwt/AquaLoRA.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-feng24k, title = {{A}qua{L}o{RA}: Toward White-box Protection for Customized Stable Diffusion Models via Watermark {L}o{RA}}, author = {Feng, Weitao and Zhou, Wenbo and He, Jiyan and Zhang, Jie and Wei, Tianyi and Li, Guanlin and Zhang, Tianwei and Zhang, Weiming and Yu, Nenghai}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {13423--13444}, year = {2024}, editor = {Salakhutdinov, Ruslan and Kolter, Zico and Heller, Katherine and Weller, Adrian and Oliver, Nuria and Scarlett, Jonathan and Berkenkamp, Felix}, volume = {235}, series = {Proceedings of Machine Learning Research}, month = {21--27 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v235/main/assets/feng24k/feng24k.pdf}, url = {https://proceedings.mlr.press/v235/feng24k.html}, abstract = {Diffusion models have achieved remarkable success in generating high-quality images. Recently, the open-source models represented by Stable Diffusion (SD) are thriving and are accessible for customization, giving rise to a vibrant community of creators and enthusiasts. However, the widespread availability of customized SD models has led to copyright concerns, like unauthorized model distribution and unconsented commercial use. To address it, recent works aim to let SD models output watermarked content for post-hoc forensics. Unfortunately, none of them can achieve the challenging white-box protection, wherein the malicious user can easily remove or replace the watermarking module to fail the subsequent verification. For this, we propose AquaLoRA as the first implementation under this scenario. Briefly, we merge watermark information into the U-Net of Stable Diffusion Models via a watermark LowRank Adaptation (LoRA) module in a two-stage manner. For watermark LoRA module, we devise a scaling matrix to achieve flexible message updates without retraining. To guarantee fidelity, we design Prior Preserving Fine-Tuning (PPFT) to ensure watermark learning with minimal impacts on model distribution, validated by proofs. Finally, we conduct extensive experiments and ablation studies to verify our design. Our code is available at github.com/Georgefwt/AquaLoRA.} }
Endnote
%0 Conference Paper %T AquaLoRA: Toward White-box Protection for Customized Stable Diffusion Models via Watermark LoRA %A Weitao Feng %A Wenbo Zhou %A Jiyan He %A Jie Zhang %A Tianyi Wei %A Guanlin Li %A Tianwei Zhang %A Weiming Zhang %A Nenghai Yu %B Proceedings of the 41st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2024 %E Ruslan Salakhutdinov %E Zico Kolter %E Katherine Heller %E Adrian Weller %E Nuria Oliver %E Jonathan Scarlett %E Felix Berkenkamp %F pmlr-v235-feng24k %I PMLR %P 13423--13444 %U https://proceedings.mlr.press/v235/feng24k.html %V 235 %X Diffusion models have achieved remarkable success in generating high-quality images. Recently, the open-source models represented by Stable Diffusion (SD) are thriving and are accessible for customization, giving rise to a vibrant community of creators and enthusiasts. However, the widespread availability of customized SD models has led to copyright concerns, like unauthorized model distribution and unconsented commercial use. To address it, recent works aim to let SD models output watermarked content for post-hoc forensics. Unfortunately, none of them can achieve the challenging white-box protection, wherein the malicious user can easily remove or replace the watermarking module to fail the subsequent verification. For this, we propose AquaLoRA as the first implementation under this scenario. Briefly, we merge watermark information into the U-Net of Stable Diffusion Models via a watermark LowRank Adaptation (LoRA) module in a two-stage manner. For watermark LoRA module, we devise a scaling matrix to achieve flexible message updates without retraining. To guarantee fidelity, we design Prior Preserving Fine-Tuning (PPFT) to ensure watermark learning with minimal impacts on model distribution, validated by proofs. Finally, we conduct extensive experiments and ablation studies to verify our design. Our code is available at github.com/Georgefwt/AquaLoRA.
APA
Feng, W., Zhou, W., He, J., Zhang, J., Wei, T., Li, G., Zhang, T., Zhang, W. & Yu, N.. (2024). AquaLoRA: Toward White-box Protection for Customized Stable Diffusion Models via Watermark LoRA. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:13423-13444 Available from https://proceedings.mlr.press/v235/feng24k.html.

Related Material