CEDAR
From DataMining
CEDAR: Conversational Entity Detector And Renderer
Most information search is driven by user queries. We explore a "push" system where the computer attempts to generate the queries. In CEDAR (Conversational Entity Detector And Renderer), as two people are speaking (or skyping), the computer listens and continually displays potentially relevant text or images. Relevant information could be drawn from both personal and public sources: my calendar and address book, Facebook (and friends), Flickr, e-mails, desktop searches, Wikipedia, IMDB, gapminder, and Google query results. Information presented could include mousable images, summaries of extracted facts, and interaction graphs of related people or concepts as obtained from friends lists. We have built a prototype system.
Main Page for the University of Pennsylvania Datamining group
