自然语言水印是一种为自然语言文档特别设计的数字版权管理技术。虽然已经有不少精心设计的自然语言水印算法,但是对它们的嵌入质量和安全性只有简单的说明。本文中,我们提出了一个自然语言水印嵌入安全性的评估方案,定义了理想不可检测性,并提出了构造针对特定检测器的不可检测性的交互证明系统。本文也对一些实际系统按此方案作了评价。
Natural language watermarking (NLW) is a kind of digital rights management (DRM) techniques specially designed for natural language documents. Though much work has been done on designing elaborate watermarking algorithms, security or quality of these schemes is only roughly discussed. In this paper, we propose an evaluating scheme for embedding security of NLW, define ideal undetectability, and design ways to construct an inter-proof system for undetectability against specific detector. We also apply this evaluating scheme to some practical NLW systems.