IBM's Watson on Jeopardy: Round 1 ends in a tie
IBM's Watson landed a tie in the first round of Jeopardy's man-vs.-machine challenge Monday night.
Watson, IBM's supercomputer, and Brad Rutter each ended Day 1 with $5,000 in winnings, while Ken Jennings, the other human competitor, came out with $2,000.
Jennings -- who had previously won 74 consecutive Jeopardy matches -- defeated Watson in the practice match between the three players leading up to the three-round faceoff.
Round 2 is Tuesday night, and the final round airs Wednesday.
While Watson proved competitive, there were some missteps.
At one point during on Monday's episode, Watson gave a wrong answer, saying "What is 1920s,” after Jennings had just given the same incorrect response.
Host Alex Trebek responded with "No, Ken already said that."
Watson, named after IBM founder Thomas J. Watson, started off Monday's episode in dominating fashion with $5,200 on the board by the first commercial break, while Rutter accumulated $1,000 and Jennings rung up $200.
The first night was a huge learning experience for the engineers who've spent years working on building Watson for this very match, said Steve Canepa, IBM's general manager of global Media and the entertainment industry.
"I've seen a lot of the the videos we've put together about the making of Watson and things like that, but I hadn't actually seen it in action until tonight," Canepa said Monday after the first night of competition.
"When Watson repeated that answer, to the general public it was probably pretty funny. But Watson only takes his input from the question board so the fact that somebody else gave the same answer already doesn't factor to into Watson says. He can't hear what the other players are saying, but maybe that's a feature we can add in the future."
Offering the same response as Jennings also shows just how smart Watson is, he said.
"There is obviously some form of logic that was very similar to that of the human player tackling that problem, and that is fascinating to me," Canepa said. "The ability to sort through what is a couple hundred million pages of information in a very short amount of time, all the data that we create in blogs and tweets and articles and all of that unstructured text on the Internet -- to be able to find the relationship between words so quickly is what the point of all this is."
All the data stored in Watson is acquired from the Internet, as well as books and journals, though for competition, Watson is disconnected from the Web, he said.
IBM is looking to change the way computers, and people, search and learn using computers, Canepa said, and Watson is searching information as humans create it, not just by data put into rows and columns as has been done before.
"We'll see what happens over the next two nights, but I'm not overly focused on the win or loss myself," he said. "In Chess, as finite as it is, there's a finite number of moves. But in this there are an infinite number of questions that can be asked and with all the puns and ways there are to ask a question.
"I'm really focused on the many real-life situations for this ability to be able to dive into unstructured data and make sense of it. The kind of search we do on a search engine today is much more keyword oriented and this is way beyond that ... If we can search with intelligence, it could open up all sorts of new fields and possibilities."
Photo: Alex Trebek, Ken Jennings, IBM's Watson, and Brad Rutter. Credit: Associated Press.