Activity-aware computing

Activity inference is a popular research topic today, blending ideas from artificial intelligence, mobile and ubiquitous computing, and user modeling. While it can enable many promising applications, these applications will not be realized unless inference is accurate enough, and user privacy and corporate confidentiality are adequately protected.

This talk will describe some of the opportunities and challenges in activity-aware computing, and the past and ongoing research at PARC toward this goal, including user modeling from heterogeneous mobile phone data sources, the utilization of publicly-available time-use data, and a better understanding of the privacy threats that arise from coarse location data.