ITL's Information Access Division (IAD) hosted the third annual Topic Detection and Tracking (TDT) Evaluation Workshop on November 16 and 17, 2000.  The workshop involved 30 academic and corporate researchers who participated in the NIST administered TDT evaluation that occurred during the fall of 2000.  The workshop followed the NIST Text Retrieval Conference (TREC), a related conference in the text retrieval research field, also hosted by IAD.

The TDT program develops technologies that search, organize and structure news-oriented textual materials from a variety of broadcast news media in both the English and Mandarin languages.  The research-driven program uses controlled laboratory simulations of hypothetical systems to test the efficacy of potential technologies to access the continuously flowing information that is available from news producing entities.

In earlier studies reported at last year's TDT meeting, research had shown that it was possible to track and organize events in news data even though the data was multilingual. The TV and radio broadcasts were transcribed using automatic speech recognition, and the Mandarin text was converted to English using COTS Mandarin to English software.

The focus of TDT changed in 2000 to development of core techniques that organize news data, and to bring more of the real world into the evaluation by requiring TDT systems to operate on broadcast news audio data that has been transcribed, translated and segmented into stories without human intervention.