arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2406.02603cs.CRcs.LG
收藏

在水印密钥碰撞下,无失真水印并非真正的无失真

Distortion-free Watermarks are not Truly Distortion-free under Watermark Key Collisions

University of Maryland(马里兰大学);University of Waterloo(滑铁卢大学)

Yihan Wu, Ruibo Chen, Zhengmian Hu, Yanshuo Chen, Junfeng Guo, Hongyang Zhang, Heng Huang

AI总结:

本文研究了无失真水印在密钥碰撞下的分布偏差问题,指出完美无失真水印不可能实现,并提出了beta-watermark以有效减少密钥碰撞引起的分布偏差。

摘要:

语言模型(LM)水印技术通过用水印密钥作为随机种子,将随机采样过程替换为伪随机采样,从而在LM生成的内容中注入统计信号。在这些统计水印方法中,无失真水印尤为重要,因为它们在不损害生成质量的情况下将水印嵌入到LM生成的内容中。然而,与真随机采样相比,伪随机采样的一个显著局限在于,在相同的水印密钥下(即密钥碰撞),伪随机采样的结果会表现出相关性。这一局限可能会破坏无失真特性。我们的研究表明,由于水印密钥的可用性有限,密钥碰撞是不可避免的,并且现有的无失真水印在存在密钥碰撞时,表现出向原始LM分布的显著分布偏差。此外,实现完美的无失真水印是不可能的,因为在密钥碰撞下无法嵌入任何统计信号。为了减少密钥碰撞引起的分布偏差,我们引入了一种新的无失真水印家族——beta-watermark。实验结果支持beta-watermark能够有效减少密钥碰撞下的分布偏差。

英文摘要:

Language model (LM) watermarking techniques inject a statistical signal into LM-generated content by substituting the random sampling process with pseudo-random sampling, using watermark keys as the random seed. Among these statistical watermarking approaches, distortion-free watermarks are particularly crucial because they embed watermarks into LM-generated content without compromising generation quality. However, one notable limitation of pseudo-random sampling compared to true-random sampling is that, under the same watermark keys (i.e., key collision), the results of pseudo-random sampling exhibit correlations. This limitation could potentially undermine the distortion-free property. Our studies reveal that key collisions are inevitable due to the limited availability of watermark keys, and existing distortion-free watermarks exhibit a significant distribution bias toward the original LM distribution in the presence of key collisions. Moreover, achieving a perfect distortion-free watermark is impossible as no statistical signal can be embedded under key collisions. To reduce the distribution bias caused by key collisions, we introduce a new family of distortion-free watermarks--beta-watermark. Experimental results support that the beta-watermark can effectively reduce the distribution bias under key collisions.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑