IBM’s Watson and the top two all-time Jeopardy! contestants. Tomorrow, the final episode will air. Since I don’t have a television, I’m forced to see the results after-the-fact by browsing through news stories on the Internet.
Apparently, Watson won the round today. However, the machine missed the final question in what was seemingly an obvious answer. Therein lies the rub. What is obvious to the human brain is oblique to a machine dumbly crunching data, searching for patterns.
I wasn’t very interested in this project until I watched the PBS NOVA episode, ‘The Smartest Machine on Earth.’ Watch it. What you’ll see is how far the programmers behind this effort have come—by painstakingly tweaking and refining algorithms—in teaching a machine to rapidly interpret complex clues. The machine learns from its mistakes.
I could go on and on, speculating about what this portends for the future of Artificial Intelligence. But I won’t. You can find that elsewhere. Suffice it to say that this is an impressive demonstration of where we are heading. I think Watson will win the contest.
I’m sure I’m not alone in thinking that this effort (and like-minded endeavors) will soon transform our lives. We’re heading towards a revolution in computer-based analysis and diagnosis. Soon, computers will capably answer complex, layered questions with unmatched speed and accuracy. Machines will be able to sift through vast pools of data to match, say, our singular health symptoms with a short list of likely causes and potential treatments—taking into account all of the most-recently published literature on the planet. Can your doctor do that?
Once machines master answering complex questions, what’s the next step? I suppose we’ll have to start teaching machines how to ask questions.