arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2211.09717cs.CYcs.CR
收藏

UPTON:通过数据投毒防止公开发布文本导致作者身份泄露

UPTON: Preventing Authorship Leakage from Public Text Release via Data Poisoning

Ziyao Wang, Thai Le, Dongwon Lee

AI总结:

针对作者在公开作品被用于训练作者身份归属模型后难以匿名写作的问题,提出数据投毒方案 UPTON,通过黑盒方法削弱作者身份特征使文本不可学习,在四个数据集上将归属准确率降至约 35% 同时保持文本可读,且对已训练模型有效。

摘要:

设想这样一个场景:一位作者(如活动家、举报人)拥有大量公开作品,在攻击者可能已经基于包括其作品在内的公开作品构建了作者身份归属(AA)模型的情况下,希望进行“匿名”写作。为实现其愿望,我们提出一个问题:“能否使公开发布的作品 T 无法被归属,从而使在 T 上训练的 AA 模型无法很好地归属其作者身份?”针对此问题,我们提出了一种新颖的解决方案 UPTON,它利用黑盒数据投毒方法削弱训练样本中的作者身份特征,使发布的文本变得不可学习。这与以往混淆工作不同——例如修改测试样本的对抗攻击,或仅在触发词出现时改变模型输出的后门工作。使用四个作者身份数据集(IMDb10、IMDb64、Enron 和 WJO),我们给出了实证验证,其中 UPTON 成功将 AA 模型的准确率降至不切实际的低水平(约 35%),同时保持文本仍可读(语义相似度 >0.9)。UPTON 对已在作者可用干净作品上训练的 AA 模型仍然有效。

英文摘要:

Consider a scenario where an author-e.g., activist, whistle-blower, with many public writings wishes to write "anonymously" when attackers may have already built an authorship attribution (AA) model based off of public writings including those of the author. To enable her wish, we ask a question "Can one make the publicly released writings, T, unattributable so that AA models trained on T cannot attribute its authorship well?" Toward this question, we present a novel solution, UPTON, that exploits black-box data poisoning methods to weaken the authorship features in training samples and make released texts unlearnable. It is different from previous obfuscation works-e.g., adversarial attacks that modify test samples or backdoor works that only change the model outputs when triggering words occur. Using four authorship datasets (IMDb10, IMDb64, Enron, and WJO), we present empirical validation where UPTON successfully downgrades the accuracy of AA models to the impractical level (~35%) while keeping texts still readable (semantic similarity>0.9). UPTON remains effective to AA models that are already trained on available clean writings of authors.

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。

↑