PO Box 2099 Mabank, TX 75147 fuzzy@lazytoad.com
A recent offshoot of the work on video retrieval and earlier work on the Tipster project is research into indexing the World Wide Web with the Lycos/Pursuit system. Lycos provides probabilistic retrieval of abstracts automatically collected from the World Wide Web.
The indexing work builds on CMU's participation in ARPA's TIPSTER Data Extraction project (in partnership with General Electric's Corporate Research and Development division). The joint GE-CMU Shogun data extraction system had the highest overall score of any group in ARPA's MUC-5 message understanding conference, and CMU's Textract system (built in cooperation with Nippon Telephone and Telegraph's Data division) scored the highest combined recall-precision score of any system evaluated by ARPA at MUC-5.
See also Biography of Michael Mauldin.