Adversarial Authorship Attribution for Deobfuscation

Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics Volume 1: Long Papers, pages 7372 - 7384 May 22-27, 2022 c⃝2022 Association for Computational Linguistics A Girl Has A Name, And It’s ...∗ Adversarial Authorship Attribution for Deobfuscation† Wanyue Zhai University of California, Davis Jonathan Rusert University of Iowa Zubair Shafiq University of California, Davis Padmini Srinivasan University of Iowa Abstract Recent advances in natural language process- ing have enabled powerful privacy-invasive au- thorship attribution. To counter authorship at- tribution, researchers have proposed a variety of rule-based and learning-based text obfusca- tion approaches. However, existing authorship obfuscation approaches do not consider the ad- versarial threat model. Specifically, they are not evaluated against adversarially trained au- thorship attributors that are aware of potential obfuscation. To fill this gap, we investigate the problem of adversarial authorship attribu- tion for deobfuscation. We show that adver- sarially trained authorship attributors are able to degrade the effectiveness of existing obfus- cators from 20-30% to 5-10%. We also eval- uate the effectiveness of adversarial training when the attributor makes incorrect assump- tions about whether and which obfuscator was used. While there is a a clear degradation in attribution accuracy, it is noteworthy that this degradation is still at or above the attribution accuracy of the attributor that is not adversar- ially trained at all. Our results underline the need for stronger obfuscation approaches that are resistant to deobfuscation. 1 Introduction Recent advances in natural language processing have enabled powerful attribution systems1 that are capable of inferring author identity by ana- lyzing text style alone (Abbasi and Chen, 2008; Narayanan et al., 2012; Overdorf and Greenstadt, 2016; Stolerman et al., 2013; Ruder et al., 2016). There have been several recent attempts to attribute the authorship of anonymously published text using ∗ This paper is third in the series. See (Mahmood et al., 2019) and (Mahmood et al., 2020) for the first two papers. †Our code and data are available at: https://github. com/reginazhai/Authorship-Deobfuscation 1https://www.eff.org/deeplinks/2013/06/internet-and- surveillance-UN-makes-the-connection such advanced authorship attribution approaches.2 This poses a serious threat to privacy-conscious individuals, especially human rights activists and journalists who seek anonymity for safety. Researchers have started to explore text obfusca- tion as a countermeasure to evade privacy-invasive authorship attribution. Anonymouth (McDonald et al., 2012; Brennan et al., 2012) was proposed to identify words or phrases that are most revealing of author identity so that these could be manually changed by users seeking anonymity. Since it can be challenging for users to manually make such changes, follow up work proposed rule-based text obfuscators that can automatically manipulate cer- tain text features (e.g., spelling or synonym) (Mc- Donald et al., 2013; Almishari et al., 2014; Keswani et al., 2016; Karadzhov et al., 2017; Castro-Castro et al., 2017; Mansoorizadeh et al., 2016; Kacmar- cik and Gamon, 2006; Kingma and Welling, 2018). Since then more sophisticated learning-based text obfuscators have been proposed that automatically manipulate text to evade state-of-the-art author- ship attribution approaches (Karadzhov et al., 2017; Shetty et al., 2018; Li et …

Content truncated to 500 words