As I’m sure you’ve heard by now, yesterday’s episode of Jeopardy had the first ever man v. machine match in Jeopardy history, with IBM ‘s supercomputer Watson taking on the best contestants in Jeopardy history–Ken Jennings (the guy who won 74 games in a row) and Brad Rutter (the biggest all-time money winner and undefeated in his 20 games).
As a SymSys student focusing in Natural Language, from the minute I first heard about Watson, I was wildly excited by what it meant that IBM had built a question-answering machine sophisticated enough to play Jeopardy and be competitive with the best contestants. We’ve been talking a lot about Watson in my natural language processing class this quarter, and even though I still know very little about the field, I know enough to be amazed at what IBM’s been able to do so far.
This is why it’s a bit disappointing to read commentary on the match like this blog post for the New York Times by Alessandra Stanley, which gives IBM little credit for the four-plus years of work that went into refining the program. She chalks up any success to Watson’s “2,800 computers,” as if to say that computers can do anything if you have enough of them. To be sure, there are aspects of Watson that make it hard to see the supercomputer as the breakthrough in artificial intelligence to end all breakthroughs. (It might have an advantage in time to process questions/buzz in, but it lacks a little common sense. It might not be sophisticated enough to decipher all the wordplay that’s characteristic of Jeopardy clues. And since it’s built to decipher and answer Jeopardy clues, which are generally different from how most people ask questions, adapting it to respond to a more general question-answer system might be extremely difficult.) But there’s still more to be impressed by than dismissive about.
And even if you could care less about what Watson actually means for the future of artificial intelligence and natural language processing, it’s still pretty awesome to see a machine play Jeopardy. It’s on again tonight and tomorrow night, take a study break and watch some Jeopardy! You might be watching the day that the machines started to take over the world.