Written language contains stylistic cues that can be exploited to automatically infer a variety of potentially sensitive author information. Adversarial stylometry intends to attack such models by rewriting an author's text. Our research proposes several components to facilitate deployment of these adversarial attacks in the wild, where neither data nor target models are accessible. We introduce a transformer-based extension of a lexical replacement attack, and show it achieves high transferability when trained on a weakly labeled corpus -- decreasing target model performance below chance. While not completely inconspicuous, our more successful attacks also prove notably less detectable by humans. Our framework therefore provides a promising direction for future privacy-preserving adversarial attacks.
翻译:书写语言包含可以用来自动推断各种潜在敏感作者信息的文体提示。 反逆性tylologization 打算通过重写作者的文字来攻击这些模型。 我们的研究提出了几个组成部分, 以便利在野外部署这些对抗性攻击, 因为在野外既无法获得数据, 也无法获得目标模型。 我们引入了基于变压器的替换攻击扩展, 并表明在使用标签不高的立体时, 它可以实现高可转移性, 也就是降低目标模型的概率。 虽然并非完全不明显, 我们更成功的攻击也证明人类难以察觉。 因此, 我们的框架为未来保护隐私的对抗性攻击提供了一个有希望的方向。