Mobile Application for Sensing, Inferring, and Explaining Availability

Anind K. Dey, Computer Science-Human Computer Interaction Institute

Staying aware of others is an established need that people have. Between close friends and family members, this can help people feel a greater sense of connectedness as each person goes about their daily activities. Between coworkers, this can inform people of the most available times to contact them. In this project, we will build a mobile application that leverages sensors on the phone, such as the GPS sensor, Wi-Fi, accelerometer, microphone, and the user's calendar to infer the user's contexts for her location, physical activity, social activity (through sound), and schedule, to ultimately infer her availability to her contacts. Some of the inference would be made through machine learning techniques (e.g., learned decision trees, naive Bayes classifier).

Due to the many factors and inference mechanisms the application uses, it will be difficult for users to understand how it determines their availability. Our novel research contribution will be the automatic provision of explanations to users so that they can better understand the application, trust it more, and use it more effectively. In this project, we will design, implement, and deploy the social-awareness mobile application to several users, and eventually conduct a user study on its effectiveness.

Skills needed / desired:

• Android programming experience
• Machine Learning experience
• Database programming with MySQL
• Server side programming (PHP or J2EE)