This Supercomputer Defeated Human Champions of a TV Game Show in 2011
It was another one of those big-thinking days at IBM today, as the supercomputer Watson–which has been prepping for a televised matchup against two human champions from the TV game show “Jeopardy”–won a practice round before a room full of reporters today.
As ZDNet reports, Watson won the round with $4,400, while Ken Jennings had $3,400 and Brad Rutter brought in $1,200.
The game has been in the planning stages for years, and has been written about several times. The New York Times covered IBM’s work in a big story in 2009.
The whole point of teaching a computer to play “Jeopardy” lies in the complex computing work that’s required to make a machine understand natural human language and detect the same subtle cues of human speech that humans learn to understand over the years. “Jeopardy” questions can involve clever turns of phrases, riddles and other tricks of speech that can confuse a computer in ways that a game of chess won’t. Computers have already defeated humans at chess, you’ll recall, and it was an IBM computer that did it.
After the “Jeopardy” match, the human players said Watson had one distinct advantage: It doesn’t get psyched out. If another player wins a string of questions, it doesn’t suffer from the emotional response of losing confidence.
I don’t usually watch “Jeopardy,” or any game show for that matter. But I’m looking forward to seeing how the real game turns out.
Below is a rough video by ZDNet’s Larry Dignan, who attended the round.