So there's this awesome competition going down...
It's a Jeopardy (answer-question trivia) match with a couple of human megachampions and Watson, a huge computer disconnected from the Internet and designed to answer things closer to how people do.
The game has already been going and tomorrow is the last day but this article explains it nicely: http://www.wired.com/epicenter/2011/...tson-jeopardy/
Since I am a huge Jeopardy nerd and very interested in cognition and artificial intelligence, I've been going crazy over this.
It's been interesting to actually watch this now and see the differences in the ways that Ken, Brad, and Watson respond.
This is a good example of a big difference (though it contains spoilers for those who haven't seen today's show yet): http://www.huffingtonpost.com/2011/0..._n_823795.html
There is a huge amount of scanning involved in the way the computer figures out the most likely correct response (80 trillion operations per second), but it can understand things in the weird ways that humans sometimes phrase them. I remember something on the show about Watson using associations, but I'm not sure what exactly is going on there.
Anyway, I'll post more stuff that I find relevant to this as I go along and frame exactly what my question is from there. I just got excited. For now, posting about how cool this is, is reason enough to post :P