Many modern machine learning applications involve sensitive correlated data, such private information on users connected together in a social network, and measurements of physical activity of a single user across time. However, the current standard of privacy in machine learning, differential privacy, cannot adequately address privacy issues in this kind of data. This work looks at a recent generalization of differential privacy, called Pufferfish, that can be used to address privacy in correlated data. The main challenge in applying Pufferfish to correlated data problems is the lack of suitable mechanisms. In this paper, we provide a general mechanism, called the Wasserstein Mechanism, which applies to any Pufferfish framework. Since the Wasserstein Mechanism may be computationally inefficient, we provide an additional mechanism, called Markov Quilt Mechanism, that applies to some practical cases such as physical activity measurements across time, and is computationally efficient.

本文介绍了一种名为Pufferfish的通用差分隐私方法，并提供了Wasserstein机制来适用于任何Pufferfish框架以解决相关数据的隐私问题，同时也提供了计算效率更高的机制来解决某些实际问题，实验结果证明了这种方法在两个不同领域的合成和真实数据中提供了隐私和实用性。

相关数据的河豚隐私机制