Watermarking Method
Watermarking methods embed hidden identifiers into digital data (images, text, models) to assert ownership and detect unauthorized modifications or usage. Current research focuses on developing robust watermarking techniques for various data types, including images (using diffusion models, Swin Transformers, and encoder-decoder architectures), language models (employing latent space manipulation and post-hoc insertion), and even tabular datasets. These advancements are crucial for protecting intellectual property, ensuring data provenance, and mitigating the risks associated with the increasing use of AI-generated content in diverse applications.
Papers
Shallow Diffuse: Robust and Invisible Watermarking through Low-Dimensional Subspaces in Diffusion Models
Wenda Li, Huijie Zhang, Qing Qu
FreqMark: Invisible Image Watermarking via Frequency Based Optimization in Latent Space
Yiyang Guo, Ruizhe Li, Mude Hui, Hanzhong Guo, Chen Zhang, Chuangjian Cai, Le Wan, Shangfei Wang