复现PATE-GAN的艰难追求：基准测试、审计与调试
The Elusive Pursuit of Reproducing PATE-GAN: Benchmarking, Auditing, Debugging
AI总结：
本文对六个开源PATE-GAN实现进行基准测试与隐私审计，发现无一能复现原始论文的效用性能，且均存在隐私泄露，并定位了19个隐私违规和5个其他错误。
摘要：
由差分隐私（DP）生成模型创建的合成数据越来越多地被用于现实场景中。在此背景下，PATE-GAN已成为最流行的算法之一，它将生成对抗网络（GANs）与PATE（教师集成的私有聚合）的私有训练方法相结合。在本文中，我们着手复现原始PATE-GAN论文中的效用评估，比较可用的实现，并进行隐私审计。更准确地说，我们分析并基准测试了六个开源的PATE-GAN实现，其中包括三个由（部分）原始作者提供的实现。首先，我们揭示了架构上的偏差，并通过实验证明，没有一个实现能够复现原始论文中所报告的效用性能。然后，我们提出了一项深入的隐私评估，其中包括DP审计，并表明所有实现泄露的隐私都超出了预期。此外，我们在这六个开源实现中发现了19个隐私违规和5个其他错误。最后，我们的代码库可从以下地址获取：https://github.com/spalabucr/pategan-audit。
英文摘要：
Synthetic data created by differentially private (DP) generative models is increasingly used in real-world settings. In this context, PATE-GAN has emerged as one of the most popular algorithms, combining Generative Adversarial Networks (GANs) with the private training approach of PATE (Private Aggregation of Teacher Ensembles). In this paper, we set out to reproduce the utility evaluation from the original PATE-GAN paper, compare available implementations, and conduct a privacy audit. More precisely, we analyze and benchmark six open-source PATE-GAN implementations, including three by (a subset of) the original authors. First, we shed light on architecture deviations and empirically demonstrate that none reproduce the utility performance reported in the original paper. We then present an in-depth privacy evaluation, which includes DP auditing, and show that all implementations leak more privacy than intended. Furthermore, we uncover 19 privacy violations and 5 other bugs in these six open-source implementations. Lastly, our codebase is available from: https://github.com/spalabucr/pategan-audit.
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。
