Carnegie Mellon University

Eric Nyberg

Bookmark and ShareTweet this storyShare this story on FacebookEmail this story with a friendSubscribe to Homepage Story RSS FeedArchivesSubmit a Story

This is Jeopardy!

Jeopardy!

Computers have long been battling human chess players. But taking on natural-language questions posed by humans — a la the game show Jeopardy! — involves a whole other level of computer wizardry.

Leave it to Carnegie Mellon professors and students, who are collaborating with IBM in order to develop a Question Answering (QA) system that can compete with humans on Jeopardy!

Not only will the system have to quickly recall knowledge from an array of topics, it will also have to evaluate its confidence level to determine whether to answer immediately or pass on a particular question.

"The Jeopardy! problem is an important challenge for the field, because systems have to respond much more quickly than they ever have in the past," explained Eric Nyberg, a professor at Carnegie Mellon's Language Technologies Institute (LTI) which is part of the School of Computer Science. "It requires forms of reasoning that are quite sophisticated — there's lots of use of metaphors, rhyming, puns and puzzles."

The field of QA initially focused on factoid questions, which only have one simple answer such as a date or a time. Now researchers at LTI are pioneering more complicated algorithms that will analyze a user's question and understand the underlying meaning of what they're asking, rather than just deciphering the words they're using.

"Carnegie Mellon is a place where the research really focuses on real-world applications," said Nyberg. "We're really trying to advance the state-of-the-art in terms of how quickly you can adapt Question Answering technology to a new business problem or a new domain."

In the future, Nyberg would like to expand his research to make the dialogue between a human and the system more natural, so it seems like there is human intelligence behind the QA system. For example, a conversational system will reply to users when their questions are too general and prompt them to narrow the focus.

With the cross-disciplinary nature at Carnegie Mellon, the possibilities are endless.

"I've been involved in a variety of research efforts that span different departments, different units, even different colleges within the university ... that's one of the unique strengths of Carnegie Mellon."

Related Links: Watch Video  |  Read IBM Release  |  School of Computer Science  |  Language Technologies Institute  |  Eric Nyberg