So IBM's new purpose-built supercomputer WATSON has been competing on Jeopardy the last 2 days. While it looked like its human opponents, Jeopardy veterans, Ken Jennings and Brad Rutter were starting to learn how to grab a few answers from him towards the end of the first day of competition, that wasn't how it went today. The biggest problem here is that I don't think its a fair fight. Watson's ability to ring in is just faster than the 2 people. It seemed like MANY of the answers WATSON got right KJ and/or BR actually knew, but were beat on the right. I just don't think that's an accurate test of Jeopardy prowess. Yes I know reaction time is part of it, but WATSON's is the same each time. The end result was WATSON wiping the floor with its opponents.
That brings us to final Jeopardy, category of "U.S. Cities".
"Its largest airport is named for a World War II hero, its second largest for a World War II battle."
(Chicago)
What really annoyed Wifey and I was the computer's final Jeopardy behavior. I find it very hard to believe that it would risk only $924 (or whatever it was) for such a general category (something its going to better at with its huge database) when earlier it risked $6K on a Double-Jeopardy question in a category about hedgehogs...
Its response was "Toronto", not even a US city. Was it programmed to answer wrong when it got to a certain point ahead of its competition?
http://www.geekosystem.com/watson-robot-jeopardy-tie/
further reading -> http://thenumerati.net/index.cfm?postID=726
&
http://artsbeat.blogs.nytimes.com/2011/02/15/watson-dominates-jeopardy-but-stumbles-over-geography/
Suggestion for further WATSON challenges below?
http://www.thehighdefinite.com/2011/02/culturally-biased/ |
No comments:
Post a Comment