Jeopardy and Watson

Post Reply New Topic

There's got to be a large number of us who enjoy Jeopardy and trivia in general. What did you guys think of the computer Watson, built by IBM to be able to parse natural language and idiom? Spoiler alert: Watson pretty much wiped the floor with Ken Jennings and Brad Rutter, the two best Jeopardy players ever. I found myself rooting for humanity and cursing Watson whenever it inevitably found the Daily Double. But I also found it fascinating that it got answers wrong on occasion and struggled with certain categories. It illustrated both how sophisticated the computer was and how far it had to go to match the natural intuition of the human brain.

Just before these shows, there was a show (novaScience?, or maybe that thing David Pogue is doing) on PBS that did a segment on Watson. My wife and I (who are both old Comp Sci majors) were talking about this right now over breakfast. Our opinion is that they have a long way to go, but it is an interesting question to think about.

I think at some point that they should create a portal that lets us ask questions, and then give feedback on the accuracy of the answer. Some sort of learning component would be needed, in my opinion. That's how we do it.

you got to figure its basicly like asking a question to google and hitting the are you feeling lucky button. the majority of the time if your specific enought it will hit 95% of the time what you want.

Watson was cool!

in a manner similar to "the enemy of my enemy is my friend," i rooted for Watson because i always felt extremely intimidated by smart people, and it was the rare smart person that was nice to me, in terms of counting them on the digits of one hand and finding the lions' share of digits left-over. so when i saw Watson reducing the two leading [pop] cultural smarties to also-rans, i cheered. i also liked Watson's "voice" a lot, it was friendly and cheerful-sounding, in a geeky sort-of way. but i was disturbed by some truly weird outliers in its responses to a few questions that even i got right. i believe though, that the brains behind watson needed to do more work on the functional interface, in that Watson had an unfair advantage in terms of it's ability to buzz-in in just 1/100th of a second, far faster than the fastest human. i believe it is this one feature which basically allowed Watson to win. the programmers should have factored this in, maybe by designing in a delay to lessed the alacrity of Watson's responses to a level still in excess of human performance, but much closer to it. just my jejune thoughts...

The chip is mightier than the word which is mightier than the sword. Therefore the chip is mightier than the sword.

ruveyn