IBM researchers are developing Watson, a natural-language processing system that will compete against human players in a game of Jeopardy! Demonstrations of the system will take place throughout the year, and a final televised match, hosted by Alex Trebek, will be held in 2010.
Questions will be spoken aloud by Trebek, but fed into the machine as text during the competition. Watson project leader and IBM computer scientist David Ferrucci says the system divides questions into pieces and searches its databases for related knowledge and makes connections to create a result. Watson is not designed to search the Web.
Whether or not Watson beats a human competitor, the project will further research in the field of natural language processing, says University of Washington computer scientist Dan Weld, who is participating in a U.S. Defense Advanced Research Projects Agency (DARPA) project focused on developing machines that understand natural language. "DARPA's involvement will focus the research of many people at top universities and research labs to push on integrated systems that can actually read a broad array of documents," Weld says.
University of California, Berkeley computer scientist Marti Hearst says that natural-language processing researchers have made significant progress in the past 10 years. She says "pitting IBM's Watson question-answering system against the top humans in a game of Jeopardy! is a fun way to publicize and showcase this progress."
From Technology Review
View Full Article
Abstracts Copyright © 2009 Information Inc., Bethesda, Maryland, USA