Skip to main content

“How IBM Plans to Win Jeopardy!”

trlogoFor decades, humans have struggled to create machines that can extract meaning from human language.  Traditional approaches require a great deal of manual work up front to render material understandable to computer algorithms, and the ultimate goal is to make this step unnecessary.  In May’s Technology Review, David Talbot discusses how IBM hopes to advance this objective.  Expected later this year, Watson — a natural-language computer system — will play Jeopardy! (the popular TV trivia show) against human contestants.

UW CSE’s Dan Weld is quoted:  “‘I expect that this whole area will heat up significantly in the next few years.”  His group has applied to take part in a five-year DARPA research effort that is aimed at advancing the state of natural-language processing.

See the full article here.

May 29, 2009