Computer beats ‘Jeopardy!’ champions

(The Beacon)
By Andrew Nuxoll, Guest Commentary -- The Beacon
This week, a computer named Watson bested the world's best "Jeopardy!" champions in a three-round televised match. As UP's resident artificial intelligence (AI) expert, I fielded several questions from my students who were anxious to know what I thought of the whole thing.
Well I think it's marvelous, but not for the reasons that you think.
When I first heard about Watson, another famous human vs. computer match-up immediately sprang to my mind. In 1997, a computer named Deep Blue decisively defeated the world chess champion in a public match. The news captivated the world because it was the first time that a computer had really beaten the best of the best that humanity had to offer in a contest generally associated with raw intelligence.
Then, like now, people began to question how long it would be before there was a C-3PO in every household or whether we should be preparing ourselves to welcome our new silicon-based overlords. In case you are wondering, my opinion both then and now is: don't hold your breath.
Another thing that Deep Blue has in common with Watson is that both computers aren't new inventions but applications of existing AI technology. In fact, some algorithms Watson uses are part of the introductory Artificial Intelligence elective course that I teach. Watson is using the same statistics-based machine-learning technology that powers Google Search and Amazon recommendations. That, coupled with some natural language tools used for language translation and, no doubt, a hefty injection of heuristics (a fancy word AI researchers use so they don't have to say "a rule of thumb").
It's because I'm familiar with Watson's underlying technology that I'm confident Watson clones won't be taking over the world. Watson may be outstanding at "Jeopardy!" but it is completely incapable of doing anything else. It can't even play trivial pursuit without the aid of a human to reprogram it for that purpose. It certainly is not going to start building homicidal robots that look like Arnold Schwarzenegger.
Now before, I start sounding like a wet blanket, let me say that Watson demonstrates the real potential that we have today, and that potential is very great indeed. AI technology is, in many aspects, facilitating the technology revolution that continues to transform our society. Ironically, it is so pervasive and reliable that it is practically invisible.
Which brings me to the real reason I think Watson is marvelous. Across the world, there are millions of people who have been inspired by watching Watson play "Jeopardy!" Of those people, more than a few are future engineers who will one day decide to enter the field of computer science in part because of this demonstration of how wonderful this oft-invisible technology is and just how colossal its potential really is.
Andrew Nuxoll is an assistant professor of engineering. He can be contacted at nuxoll@up.edu.
Curious about what Watson got wrong?
Some nuanced questions were difficult for Watson to grasp such as:
Question: "From the Latin for end, this is where trains can also originate."
Answer: "terminus"