I never said she stole my money

There was an interesting article in the Times last week about IBM’s attempt to create a computer Jeopardy! champion, now that they have conquered the chess world.  Storing and retrieving all the information necessary is easy for a computer.  The hard part, is simply interpreting the English language as it is actually used:

Indeed, the creators of the system — which the company refers to as Watson, after the I.B.M. founder, Thomas J. Watson Sr. — said they were not yet confident their system would be able to compete successfully on the show, on which human champions typically provide correct responses 85 percent of the time.

“The big goal is to get computers to be able to converse in human terms,” said the team leader, David A. Ferrucci, an I.B.M. artificial intelligence researcher. “And we’re not there yet.”

The team is aiming not at a true thinking machine but at a new class of software that can “understand” human questions and respond to them correctly. Such a program would have enormous economic implications. Despite more than four decades of experimentation in artificial intelligence, scientists have made only modest progress until now toward building machines that can understand language and interact with humans.

The current program, “Watson” is quite hit and miss…

In a demonstration match here at the I.B.M. laboratory against two researchers recently, Watson appeared to be both aggressive and competent, but also made the occasional puzzling blunder. For example, given the statement, “Bordered by Syria and Israel, this small country is only 135 miles long and 35 miles wide,” Watson beat its human competitors by quickly answering, “What is Lebanon?”

Moments later, however, the program stumbled when it decided it had high confidence that a “sheet” was a fruit.

Here’s my favorite part of the article, though, and the reason for the title of the post:

The way to deal with such problems, Dr. Ferrucci said, is to improve the program’s ability to understand the way “Jeopardy!” clues are offered. The complexity of the challenge is underscored by the
subtlety involved in capturing the exact meaning of a spoken sentence. For example, the sentence “I never said she stole my money” can have seven different meanings depending on which word is stressed.

“We love those sentences,” Dr. Nyberg said. “Those are the ones we talk about when we’re sitting around having beers after work.”

If you are like everbody else I know, you won’t be able to resist trying that out all seven ways.

Advertisements

About Steve Greene
Professor of Political Science at NC State http://faculty.chass.ncsu.edu/shgreene

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: