Watson Jeopardy

A few of my friends on Facebook are Jeopardy freaks. They watch it daily and seem to ask me about once a week if I knew the answer the Final Jeopardy question (most of the time I do; I have a lot of useless knowledge stored in my brain. For some reason I know that 5 to the fourth power is 625; I also know that Simpsons producer Al Jean’s dad has a hardware store somewhere near Detroit). They were up in arms when the computer Watson beat everyone.

Amamzingly, a computer is faster on Google than you are.

“We have to stop Watson before he turns into Skynet!” “Watson will kill us all!” and “Watson has come back from the future to terminate us all” are actual status updates on Facebook. I say this to each of them:

Jeopardy is not chess. There aren’t near-infinite possible repercussions when you make your next move. All a computer does is store and regurgitate information. As a user, it looks like we’re creating some new website or using Illustrator to produce something that has never been seen before, but we’re not. We’re typing in words for the computer to memorize and regurgitate when asked; we’re telling it vectors to memorize and regurgitate when asked.

Jeopardy is a memorization show. Who here can memorize and regurgitate the most information the quickest?

You?

Me?

Ken Jennings?

No. Any computer anywhere can memorize things faster than you can.

We’re not asking it to be creative or come up with a new way to solve a problem. We’re not even asking it to calculate the most logical route to take (like a GPS or a computer chess opponent). We’re asking it to look up shit real quick on the internet. Or, better yet, we’re asking it to download Wikipedia and look up shit even quicker.

A computer beating a man at something like chess is an amazing feat for the computer programmer: he studied how chess champions think and figured out a way to combat that using zeroes and ones. A computer beating a man at trivia is not an amazing feat for anyone. That’s what they’re designed to do: remember stuff so we don’t have to.

Advertisements

One Response

  1. Obviously, knowledge retention is not Watson’s amazing ability. Computers have long been superior in this realm. Watson’s amazing-for-a-computer ability is its language parsing skills.

    Remember that Jeopardy questions are not always (often?) straightforward. For instance, one category on Watson’s run involved “Literary APBs”: the clue would describe a literary character as if it were describing a police suspect, and the players had to name the literary character described; this was certainly not a format Watson had been programmed to recognize. Jeopardy is not simply a game of trivial pursuit; its clues are often authored in an obfuscatory way, resembling riddles more than straightforward trivia. The Watson games were no exception.

    We’re accustomed to working with immensely knowledgeable computers that require us to request information in very specific ways in order to retrieve it. Watson, however, is an immensely knowledgeable computer that can understand even deliberately obfuscatory information requests. It is a great achievement, both more difficult and more important than Deep Blue.

    (Ironically, Deep Blue was more emblematic of the criticism you levy. Deep Blue was a brute force chess machine doing what chess machines do well: calculating potential future positions. It was fed large amounts of data regarding chess games, and given great processing power. None of this is to say that making Deep Blue was easy, but it was certainly not an astonishing leap forward in the development of artificial intelligence or computing power. It was simply the last step in a string of ever-more-powerful chess machines. Deep Blue did not succeed due to its creators’ superior creativity or chess mastery.)

    (Watson’s not necessarily that different, I guess, being similar to simpler machines/programs, but having a greater store of information and processing power.)

    (You know what, it’s all bullshit. Who cares? Computers can do anything if you spend that much time and money to make ti happen. IBM can shit out a Turing test any time they want.)

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: