标题： 扩散和自回归图像生成模型中的放射性水印 显示英文标题

标题： Radioactive Watermarks in Diffusion and Autoregressive Image Generative Models

摘要： 生成模型在图像生成领域变得越来越流行，但训练它们需要大量数据集，而收集和整理这些数据集的成本很高。 为了规避这些成本，一些方可能会通过使用生成的图像作为他们自己模型的训练数据来利用现有的模型。 一般来说，水印是一种检测生成图像未经授权使用的有价值工具。 然而，当这些图像被用于训练新模型时，只有当水印在训练过程中保持不变，并且在新训练模型的输出中仍可识别时，水印才能实现检测——这一特性被称为放射性。 我们分析了扩散模型（DMs）和图像自回归模型（IARs）生成图像中的水印的放射性。 我们发现现有的DMs水印方法无法保持放射性，因为水印在编码到潜在空间的过程中被擦除，或者在噪声-去噪过程中丢失（在潜在空间中的训练过程中）。 同时，尽管IARs最近在图像生成质量和效率方面超过了DMs，但尚未有针对IARs的放射性水印方法被提出。 为克服这一限制，我们提出了第一个专为IARs设计并考虑放射性的水印方法——从大型语言模型（LLMs）中的技术获得灵感，这些模型与IARs具有相同的自回归范式。 我们的广泛实验评估突显了我们方法在IARs中保持放射性的有效性，实现了强大的来源追踪，并防止了其生成图像的未经授权使用。 显示英文摘要

摘要： Image generative models have become increasingly popular, but training them requires large datasets that are costly to collect and curate. To circumvent these costs, some parties may exploit existing models by using the generated images as training data for their own models. In general, watermarking is a valuable tool for detecting unauthorized use of generated images. However, when these images are used to train a new model, watermarking can only enable detection if the watermark persists through training and remains identifiable in the outputs of the newly trained model - a property known as radioactivity. We analyze the radioactivity of watermarks in images generated by diffusion models (DMs) and image autoregressive models (IARs). We find that existing watermarking methods for DMs fail to retain radioactivity, as watermarks are either erased during encoding into the latent space or lost in the noising-denoising process (during the training in the latent space). Meanwhile, despite IARs having recently surpassed DMs in image generation quality and efficiency, no radioactive watermarking methods have been proposed for them. To overcome this limitation, we propose the first watermarking method tailored for IARs and with radioactivity in mind - drawing inspiration from techniques in large language models (LLMs), which share IARs' autoregressive paradigm. Our extensive experimental evaluation highlights our method's effectiveness in preserving radioactivity within IARs, enabling robust provenance tracking, and preventing unauthorized use of their generated images.