Topic Detection and Tracking Pilot Study Final Report

Topic Detection and Tracking (TDT) is a DARPA-sponsored initiative to investigate the state of the art in finding and following new events in a stream of broadcast news stories. The TDT problem consists of three major tasks: (1) segmenting a stream of data, especially recognized speech, into distinct stories; (2) identifying those news stories that are the first to discuss a new event occurring in the news; and (3) given a small number of sample news stories about an event, finding all following stories in the stream. The TDT Pilot Study ran from September 1996 through October 1997. The primary participants were DARPA, Carnegie Mellon University, Dragon Systems, and the University of Massachusetts at Amherst. This report summarizes the findings of the pilot study. The TDT work continues in a new project involving larger training and test corpora, more active participants, and a more broadly defined notion of “topic” than was used in the pilot study. 

The following individuals participated in the research reported. 

James Allan, UMass 

Brian Archibald, CMU 

Doug Beeferman, CMU 

Adam Berger, CMU 

Ralf Brown, CMU 

Jaime Carbonell, CMU 

Ira Carp, Dragon 

Bruce Croft, UMass, 

George Doddington, DARPA 

Larry Gillick, Dragon 

Alex Hauptmann, CMU 

John Lafferty, CMU 

Victor Lavrenko, UMass 

Xin Liu, CMU 

Steve Lowe, Dragon 

Paul van Mulbregt, Dragon 

Ron Papka, UMass 

Thomas Pierce, CMU 

Jay Ponte, UMass 

Mike Scudder, UMass 

Charles Wayne, DARPA 

Jon Yamron, Dragon 

Yiming Yang, CMU