IBM Supercomputer Takes On Jeopardy Stars

ibm-watson

IBM scientists Dave Ferrucci with the Watson supercomputer, which will compete against two human champions on the game show Jeopardy.

The latest test of man vs. machine is set, and should be a doozy. Back in August we wrote about IBM's plans to pit one of its Blue Gene supercomputers against human players on the "Jeopardy" game show.  Now the dates and opponents have been set. IBM announced today that its supercomputer will face the two most successful Jeopardy contestants of all time, Ken Jennings and Brad Rutter, in a series of matches that will will air on February 14, 15 and 16.

The IBM team scientists believe the computing system DeepQA, nicknamed "Watson" after the IBM founder Thomas J. Watson, will be able to understand complex questions and answer with enough precision and speed to compete effectively.

"After four years, our scientific team believes that Watson is ready for this challenge based on its ability to rapidly comprehend what the Jeopardy! clue is asking, analyze the information it has access to, come up with precise answers, and develop an accurate confidence in its response,” said Dr. David Ferrucci, the scientist leading the IBM Research team that has created Watson.

Prize of $1 Million
Watson and its human opponents will compete for a top prize of $1 million, with second place earning $300,000 and third place $200,000. Rutter and Jennings will donate 50 percent of their winnings to charity and IBM will donate 100 percent of Watson's winnings to charity.

It's not IBM's first effort to match its suerpcomputers against human opponents. In  1997, an IBM computer called “Deep Blue” defeated World Chess Champion Garry Kasparov in a famous battle of human versus machine. As in that showdown, IBM has elected to test its technology against the toughest human opposition.

Jennings broke the Jeopardy! record for the most consecutive games played by winning 74 games in a row during the 2004-2005 season, resulting in winnings of more than $2.5 million. Rutter won a record $3.25 million between his original appearance in 2002 and subsequent  "tournament of champions" matches.

Watson uses massively parallel processing to parse complex questions. The approach marries advanced machine learning and statistical techniques with the latest in natural language processing to result in human-like precision and speed. In preparation for the match, Watson played more than 50 "sparring" matches this fall against former Jeopardy Tournament of Champions contestants. Watson has also taken and passed the same Jeopardy contestant test that humans take to qualify to play on the show.

Inherent Challenges in Playing Jeopardy!
Competing on Jeopardy poses a challenge for a computing system due to the variety of subject expertise required, including a knowledge of history, literature, politics, film, pop culture and science. The clues given to contestants often involve analyzing subtle meaning, irony, riddles, and other complexities at which humans excel and computers traditionally do not. Just like human competitors, Watson will not be connected to the Internet or be able to access outside assistance.

In the video below IBM discusses natural language processing and how Watson works.

For more information, visit the project's page on the IBM website.