重参数化技巧是如何破坏差分隐私文本表示学习的
How reparametrization trick broke differentially-private text representation learning
AI总结：
本文形式化分析了几篇基于DPText的NLP论文，揭示了重参数化技巧导致其差分隐私声明失效，并提出经验性检查方法以判断DP机制实现是否违反隐私损失保证。
摘要：
随着隐私在 NLP 社区中日益受到重视，研究人员开始采用各种隐私保护方法。作为最受欢迎的隐私框架之一，差分隐私（DP）凭借其基础的理论保证，或许是最具吸引力的。尽管差分隐私的总体概念看似简单，但将其应用于 NLP 时，要正确实现却并非易事。在这篇短文中，我们从形式上分析了几篇近期提出使用 DPText 进行文本表示学习的 NLP 论文，并揭示了它们关于具备差分隐私的虚假声明。此外，我们还展示了一个简单而通用的经验性健全性检查，用于确定给定的 DP 机制实现是否几乎必然违反了隐私损失保证。我们的主要目标是提高意识，帮助社区理解将差分隐私应用于文本表示学习的潜在陷阱。
英文摘要：
As privacy gains traction in the NLP community, researchers have started adopting various approaches to privacy-preserving methods. One of the favorite privacy frameworks, differential privacy (DP), is perhaps the most compelling thanks to its fundamental theoretical guarantees. Despite the apparent simplicity of the general concept of differential privacy, it seems non-trivial to get it right when applying it to NLP. In this short paper, we formally analyze several recent NLP papers proposing text representation learning using DPText (Beigi et al., 2019a,b; Alnasser et al., 2021; Beigi et al., 2021) and reveal their false claims of being differentially private. Furthermore, we also show a simple yet general empirical sanity check to determine whether a given implementation of a DP mechanism almost certainly violates the privacy loss guarantees. Our main goal is to raise awareness and help the community understand potential pitfalls of applying differential privacy to text representation learning.
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。
