Flashgroups: Automatically Forming Collaborative Groups of Analysts around Prioritized Information

The web has experienced a huge increase in the amount of content published daily, forcing users to cope with information overload. This challenge,however, is not limited to just web content. More recently, an explosion in social content and social connections on the web have lead to a new challenge: social overload. The DoD is also faced with similar information and social overload challenges: an intelligence analyst must wade through huge amounts of possible documents, both from classified sources and from the web. Similarly, from a''social'perspective, there many be a number of different analysts working on similar or related tasks. If these analysts could collaborate effectively, each of their individual assessments could be more accurate and timely. However, it is very difficult to discover such shared context and to form the associated groups. In the proposed effort, we will develop and deploy a general tool, Flashgroup, for addressing such information and social overload challenges. In particular, our system will be able to quickly model a user's intention or context, recommend documents that are most appropriate for their current goals, and suggest groups of users, or flashgroups, who share a common interest to enable engaging and rewarding collaborations.