arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2307.13808cs.CLcs.CR
收藏

面向AI检测的条件文本生成水印技术:揭示挑战与一种语义感知水印补救方案

Watermarking Conditional Text Generation for AI Detection: Unveiling Challenges and a Semantic-Aware Watermark Remedy

University of California Riverside(加州大学河滨分校);Tianjin University(天津大学)

Yu Fu, Deyi Xiong, Yue Dong

AI总结:

针对现有水印技术损害条件文本生成性能的问题,提出一种考虑输入上下文的语义感知水印算法,在BART和Flan-T5等模型上显著提升了摘要等任务的生成质量并保持检测能力。

摘要:

为了减轻与语言模型相关的潜在风险,近期的AI检测研究提出通过随机词汇限制将水印融入机器生成的文本,并利用此信息进行检测。虽然这些水印仅导致困惑度的轻微恶化,但我们的实证调查揭示了对条件文本生成性能的显著损害。为解决此问题,我们引入了一种简单而有效的语义感知水印算法,该算法考虑了条件文本生成的特征和输入上下文。实验结果表明,我们提出的方法在摘要生成和数据到文本生成等任务中,在包括BART和Flan-T5在内的各种文本生成模型上取得了显著改进,同时保持了检测能力。

英文摘要:

To mitigate potential risks associated with language models, recent AI detection research proposes incorporating watermarks into machine-generated text through random vocabulary restrictions and utilizing this information for detection. While these watermarks only induce a slight deterioration in perplexity, our empirical investigation reveals a significant detriment to the performance of conditional text generation. To address this issue, we introduce a simple yet effective semantic-aware watermarking algorithm that considers the characteristics of conditional text generation and the input context. Experimental results demonstrate that our proposed method yields substantial improvements across various text generation models, including BART and Flan-T5, in tasks such as summarization and data-to-text generation while maintaining detection ability.

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。

↑