The Automatic Detection of Sensitive Data in Smart Homes

摘要

Smart homes are increasingly becoming popular because they make living comfortable, enjoyable, and secure. People can remotely control various aspects of their smart home environments. However, smart home appliances can pose threats to privacy. The reason is that smart appliances collect and store sensitive information, and if hackers gain access to this information, user privacy may be breached. It is difficult for users to constantly monitor and determine which data is sensitive to them and which one is not. Also, a user's identity can be leaked during sharing of information with different service providers such as health care providers and utility companies. In this paper we address one important privacy issue in smart homes, which is lack of users' control over their desired privacy. We propose a privacy decision framework which considers this problem. In this framework, active learning (machine learning) technique is used to help users detect sensitive information according to their privacy preferences.