EAGER: Collaborative Research: Towards Understanding Smartphone User Privacy: Implication, Derivation, and Protection

该项目旨在解决智能手机用户对隐私的担忧。特别是，它调查了智能手机应用程序(App)的使用可能如何重塑用户的隐私感知，以及这种重塑的含义是什么。最近有调查隐私泄露和潜在防御机制的工作。然而，到目前为止，人们对这种隐私损失的后果了解有限，特别是在智能手机用户通过许多应用程序泄露大量隐私信息的情况下。该项目旨在调查移动技术(即智能手机和智能手机应用程序)如何泄露用户的个人信息，并通过考虑用户的社会关系来确定侵犯隐私的后果。该项目有助于深入了解移动设备时代的用户隐私，并进一步开发适当的保护机制。基于不同程度的信息泄露，分析了不同层次的智能手机用户隐私，包括个人、社会和社区关系。在大规模跟踪驱动的调查和实验研究的基础上，建立了贝叶斯网络和隐马尔可夫模型等统计模型来理解用户的时间隐私泄露模式。开发数据可视化工具，实时捕获和显示不同类型隐私泄露的时空模式和汇总统计数据，帮助用户更好地洞察潜在的隐私损失。统计建模和数据可视化技术进一步使社会科学家能够研究侵犯隐私的心理或社会后果，并识别鼓励对智能手机用户隐私的关注或疏忽的因素。

This project aims to address privacy concerns of smartphone users. In particular, it investigates how the usages of the smartphone applications (apps) may reshape users' privacy perceptions and what is the implication of such reshaping. There has been recent work that investigates privacy leakage and potential defense mechanisms. However, so far there is only limited understanding on the consequences of such privacy losses, especially when large amount of privacy information leaked from smartphone users across many apps. The project seeks to investigate how the mobile technology (i.e., smartphone and smartphone apps) can reveal users' personal information and identify the consequences of privacy violations, by taking users' social relationships into consideration. The project facilitates a deep understanding of user privacy in the age of mobile devices and further develops appropriate protective mechanisms. Smartphone user privacy across different levels are analyzed including individual, social and community relationships based on different levels of information leakage. Statistical models, such as Bayesian networks and hidden Markov models, are developed to understand users' temporal privacy leakage patterns based on large-scale trace-driven investigation and experimental study. Data visualization tools are developed to capture and display the spatial-temporal patterns and summary statistics of different types of privacy leakage in real time, which helps users gain better insights on potential privacy losses. The statistical modeling and the data visualization techniques further enable the social scientists to study the psychological or social consequences of privacy violations, and identify factors encouraging attention or inattention to smartphone user privacy.