IBM's Watson Finds 'Jeopardy' Elementary
The finals of its highly publicized "Jeopardy" tournament found IBM's "Watson" computer system handily winning with a total of US$77,147 -- upwards of $50,000 more than the sums amassed by Ken Jennings and Brad Rutter, its grand champion adversaries.
Watson's $1 million grand prize will be split equally between World Vision, a community development, disaster relief and advocacy organization, and the World Community Grid, a public computing grid that tackles beneficial projects such as HIV/AIDS treatments, cancer research and affordable water purification. Both charities were designated by IBM.
Watson finished off Jennings and Rutter with what, in a human competitor, might have been termed "merciless efficiency" -- a description that seems redundant when applied to a state-of-the-art computing system.
Watson's Goofy Streak
Popular reports compared Watson to numerous computers inhabiting science fiction and popular culture, but such depictions of human-like digital intelligence tend to be painted as the good (the original "Star Trek" ship computer sultrily voiced by Majel Barrett-Roddenberry), the bad (the disturbingly even-tempered HAL from "2001: A Space Odyssey") and the ugly ("Terminator's" homicidal Skynet and its cyborg offspring).
Watson was certainly as calm, collected and clinically efficient as any of its fictional counterparts, but the system also displayed eccentric quirks and was prone to occasional goofy mistakes. What weird mojo did it use to calculate its Daily Double wagers? Toronto? Really?
Bottom line: Watson certainly appears "smart" in the sense of providing precise and typically correct answers to verbally stated complex questions, but it also seems like it could be engagingly, even playfully useful.
Student Becomes Teacher
The fact is that "Jeopardy" is as educational a program as it is entertaining. Sure, "Jeopardy" is simply a televised version of "Trivial Pursuit" on steroids, but the program also considers obscure subjects and disciplines with which many watchers are likely unfamiliar.
Plus, contradicting a world where increasingly specific sub-specialization is most amply rewarded, "Jeopardy" contestants tend to win more depending on the breadth of their knowledge rather than its depth.
That said, specialized skills clearly abound among Watson's creators; IBM and other researchers painstakingly broke down question/answer processes of language into their tiniest, most dynamic nuts and bolts (resulting in what IBM prosaically refers to as "Deep Q&A").
By doing so, the company created a computer able to understand and respond to natural language. That Watson is a system as capable of teaching as it is of being "taught" certainly evokes some irony, but that point should inform and expand the commercial applications of Watson's various technologies.
Fictional anthropomorphized computers aside, Watson certainly has real-world precedents, the foremost of which was IBM's Deep Blue system that successfully challenged World Chess Champion Garry Kasparov in a series of highly publicized 1997 matches.
However, the choice of "Jeopardy" for Watson's debut was simply brilliant from both publicity and entertainment standpoints. The matches also generated significantly higher audiences for "Jeopardy" than usual. Reportedly, the Watson contests attracted about 30 percent more viewers than "Jeopardy" episodes in February 2010, and Watson was the No. 2 ranked show in numerous major markets.
Good enough -- but was this really such a big deal? Absolutely. "Jeopardy" has a firm position in television culture and a large dedicated audience that would be naturally interested in a unique tournament.
However, the numeric impact of Watson's "Jeopardy" triumph is far greater than Deep Blue's matches against Kasparov since virtually any daily "Jeopardy" audience is far larger than those following chess tournaments, even world championship events.
The Way We Know Computers
A contributing factor to Watson's popularity is the degree to which computing technology has become commonplace in the last decade. Consider this: The year Deep Blue beat Kasparov, the World Wide Web was less than a decade old. PCs were increasingly common in businesses and homes, as was Internet connectivity (usually via dial-up modem for consumers).
Apple was teetering on bankruptcy (eventually taking a $150 million lifeline from Microsoft). "Quake" was a year old, Amazon.com, Java and the Sony PlayStation were two, and Windows 98 was a year away from launch.
Today, computers -- personal and those operating invisibly behind the scenes supporting various infrastructures -- influence, inform and impact nearly every aspect of life in the developed world.
Traditional PCs continue to inhabit homes and businesses but they are usually supported and sometimes supplanted by increasingly mobile devices including laptops, netbooks, tablets and smartphones. Perhaps most importantly, the Internet is always on, always available and almost always accessible.
What does that have to do with Watson and "Jeopardy"? In 1997, IBM's Deep Blue represented the stuff of science fiction, but it touched most deeply an audience that was technically and intellectually astute.
Watson, on the other hand, was presented to and came within the grasp and imaginations of a far larger, more mainstream audience -- folks who use and interact with complex technologies every day.
As a result, we expect that IBM Watson's "Jeopardy" appearance will remain firmly in the minds of many more people than Deep Blue ever reached, and could even be remembered as the event that fundamentally changed the relationship between people and machines.