Skip to main content [aditude-amp id="stickyleaderboard" targeting='{"env":"staging","page_type":"article","post_id":237612,"post_type":"story","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"business,","session":"B"}']

IBM’s Watson AI takes on Jeopardy’s best contestants

IBM Jeopardy Challenge

Today, for the first time ever, IBM demonstrated its Watson artificial intelligence supercomputer at a press conference at IBM Research’s Worldwide Headquarters in Yorktown Heights, New York.

[aditude-amp id="flyingcarpet" targeting='{"env":"staging","page_type":"article","post_id":237612,"post_type":"story","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"business,","session":"B"}']

IBM brought in top Jeopardy contestants Ken Jennings and Brad Rutter for a short demonstration match with Watson, in which it trounced both of the humans (we’ll have video of the match soon). Watson will make its television debut on Jeopardy in two matches across three nights, from February 14 to 16, competing against Jennings and Rutter for a $1 million prize. It’s the first time a machine has competed on Jeopardy, and I get the feeling it won’t be the last.

25 IBM Research scientists across the world toiled for four years on Watson, which is IBM’s spiritual successor to Deep Blue, the supercomputer that defeated chess grandmaster Gary Kasparov in 1997. IBM describes Watson as “an analytical computing system that specializes in natural human language and provides specific answers to complex questions at rapid speeds.”

AI Weekly

The must-read newsletter for AI and Big Data industry written by Khari Johnson, Kyle Wiggers, and Seth Colaner.

Included with VentureBeat Insider and VentureBeat VIP memberships.

Unlike Deep Blue, Watson needs to do more than just decide the next chess move, something that relies heavily on mathematical calculations. Watson instead has to interpret human language. When it comes to Jeopardy, the system has to understand the question being asked and then search its database of 200 million pages of content to determine the answer. It needs to decide when to take risks, which questions to bet on and how much to bet. IBM says the toughest part for the computer is finding and justifying the correct answer, or in other words computing a confidence that it’s right — something that comes naturally to the top Jeopardy players.

IBM says that the real world applications of Watson are far-reaching: Such a powerful and fast analytical system could save lives when it comes to health diagnosis, IBM’s primary focus at the moment, and it could also help with things like tech support and enterprise knowledge management.

When asked if the US government has shown any interest in the system, IBM would only say it’s been speaking with many interested parties. But it’s not that difficult to see how Watson could be used to help analysts at agencies like the CIA sift through vast amounts of data.

Watson is powered by 10 racks of IBM Power 750 Linux servers with 2,880 processor cores running at 80 teraflops and 15 terabytes of RAM. Deep Blue, on the other hand, ran at around 1 teraflop.

Watson is still dealing with some issues. Jeopardy host Alex Trebek pointed out that Watson confused actor Jamie Foxx with Beethoven in a question regarding his film The Soloist. IBM also remains coy about how Watson bets funds for Jeopardy’s Daily Double question — Trebek remarked that Watson made a peculiar Daily Double wager earlier in the day.

We’re waiting for IBM to put video of the test match online, which will give you the best sense of Watson’s capabilities. Watch for a post on that later today.

[aditude-amp id="medium1" targeting='{"env":"staging","page_type":"article","post_id":237612,"post_type":"story","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"business,","session":"B"}']

VentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Learn More