Recently, diffusion models (DMs) have demonstrated their advantageous potential for generative tasks. Widespread interest exists in incorporating DMs into downstream applications, such as producing or editing photorealistic images. However, practical deployment and unprecedented power of DMs raise legal issues, including copyright protection and monitoring of generated content. In this regard, watermarking has been a proven solution for copyright protection and content monitoring, but it is underexplored in the DMs literature. Specifically, DMs generate samples from longer tracks and may have newly designed multimodal structures, necessitating the modification of conventional watermarking pipelines. To this end, we conduct comprehensive analyses and derive a recipe for efficiently watermarking state-of-the-art DMs (e.g., Stable Diffusion), via training from scratch or finetuning. Our recipe is straightforward but involves empirically ablated implementation details, providing a solid foundation for future research on watermarking DMs. Our Code: https://github.com/yunqing-me/WatermarkDM.

翻译：近期，扩散模型在生成任务中展现出显著优势。将扩散模型集成至下游应用（如生成或编辑逼真图像）引发了广泛关注。然而，扩散模型的实际部署与空前能力引发了包括生成内容版权保护与监控在内的法律问题。在此背景下，水印技术已被验证为版权保护与内容监控的有效解决方案，但在扩散模型领域的研究尚不充分。具体而言，扩散模型从更长的生成轨迹中采样样本，且可能具有新型多模态结构，需对传统水印流程进行改进。为此，我们通过系统性分析，推导出对最先进的扩散模型（如稳定扩散模型）进行高效水印嵌入的完整配方——既可采用从头训练方式，也可通过微调实现。本配方设计直观，但包含经过实证分析的消融实现细节，为扩散模型水印技术的后续研究奠定了坚实基础。代码地址：https://github.com/yunqing-me/WatermarkDM