The popularity of context-aware services is improving the quality of life, while raising serious privacy issues. In order for users to receive quality service, they are at risk of leaking private information by adversaries that are possibly eavesdropping on the data and/or by the untrusted service platform selling off its data to adversaries. Game theory has been utilized as a powerful tool to achieve privacy preservation by strategically balancing the trade-off between profit (service) and cost (data leakage) for the user. However, most of the existing schemes cannot fully exploit the power of game theory, as they fail to depict the mutual relationship between any two (of the three) parties involved: user, platform, and adversary. Existing schemes are also not always able to provide specific guidance for a user to reduce the impact of the joint threats from the platform and adversary. In this paper, we design a privacy-preserving game to quantify the three parties’ concerns and capture interactions between any two of them. We also identify the best strategy for each party at a fine-grained level, i.e. specific settings, not simply binary choices. We validate the performance of our proposed game model through both a theoretical analysis and experiments.

情境感知服务的普及正在提高生活质量，同时也引发了严重的隐私问题。为了让用户获得优质服务，他们面临着被可能窃听数据的对手和/或不可信的服务平台将其数据出售给对手而泄露私人信息的风险。博弈论已被用作通过战略性地平衡用户的利润（服务）和成本（数据泄漏）之间的权衡来实现隐私保护的强大工具。然而，大多数现有方案无法充分利用博弈论的力量，因为它们未能描述所涉及的任何两方（三方中的）之间的相互关系：用户、平台和对手。现有的方案也不总是能够为用户提供具体的指导，以减少来自平台和对手的联合威胁的影响。在本文中，我们设计了一个隐私保护游戏来量化三方的关注点并捕获其中任意两个方之间的交互。我们还在细粒度级别确定了每一方的最佳策略，即特定的设置，而不是简单的二元选择。我们通过理论分析和实验验证了我们提出的博弈模型的性能。