Protecting Mobile Devices from Adversarial User by Fine-Grained Analysis of User Behavior

In recent years, mobile devices have become the mainstream medium for maintaining and processing a variety of information, including personal or sensitive information. To prevent a leakage of information to unauthorized users, a user authentication scheme that is appropriate for mobile environment is surely needed. In the meantime, it is important to guarantee sufficient usability of the mechanism so that it does not affect the user experience of the mobile device in a significant way. To pursue both usability and security perspectives of mobile devices, we propose a novel way of user authentication where fine-grained user behaviors are exploited to prohibit unauthorized access from an adversarial user. We examine behavioral patterns of daily usage of mobile devices in very fine-grained and implicit fashion. As a result, we suggest and evaluate the overall system exploiting user behaviors to separate valid user and adversarial user.