Trivia Pursuit: Analytics, Entertainment and IBM’s Watson Computer
Starting Monday, IBM pitted its supercomputer, Watson, in a three-episode Jeopardy contest against the trivia television show’s two greatest champions, Ken Jennings and Brad Rutter. With the company touting powerful analytics capabilities at play and the promising long-range implications from Watson’s showing, the world of data management is getting a unique mix of public attention and the entertainment spotlight. Here, Information Management will follow up on our preview of the match and the computer's data churning capabilities with daily reviews of important data aspects and outside commentary.
Day One – February 14
Before any questions (or answers) popped up on the show’s famous blue screens, Alex Trebek and cast took a longer look at what Big Blue had been doing to develop Watson during the past four years. The racks of servers and massive cooling station that actually make up Watson – an avatar with an IBM logo took the computer’s place between Jennings and Rutter on stage – received plenty of air-time. Maybe the lack of a domineering robot cut down on the entertainment factor, but it gave IBM a platform to explain exactly what and why it’s going through with this competition.
Then, in the game, Watson got off to a rousing start. Graphic lines spun around the avatar’s globe (“sweat,” the show called it) as it correctly pegged questions on The Beatles and won big on the Daily Double.
As the round wore on, Watson missed a few questions, in what Pervasive’s Dave Kasabian pointed to as an issue of context in language and analytics. And open source writer Marcio Saito opined on Twitter that the computer’s answering ability seemed to have more in common with pattern matching than analyzing language.
After some curious wrong answers and incredible speed and accuracy with others, Watson finished the first round tied for first with Rutter. IBM already has some outside connections in place with the technology. IBM announced prior to the televised event that eight universities will be helping develop an open architecture related to Watson’s Question Answering analytics capabilities.The computer’s long-range future rests in IBM’s ability to, as Thomas Reuters research scientist Dr. Jochen L. Leidner put it, monetize its capabilities in a “commercialization afterlife.”
Day Two – February 15
Watson buried his human competition on day two of the Jeopardy/IBM challenge. But you might not want the supercomputer to plan your next vacation.
With the speed of Watson’s natural language processing algorithms on display, the computer jumped out first with correct answers to nearly every question in the double Jeopardy round. It amassed more than three times the dollar total than second-place competitor Rutter, racking up some precise (read: $947) wager amounts along the way. However, as witnessed a few times already in the match, the answer machine produced some odd results from its HAL-like voice. Particularly, the layered final Jeopardy answer involving U.S. airports and both World Wars left the computer guessing (complete with question marks) “Toronto?”
A misstep, for sure, but as technology analyst Stephen Baker writes, it is a reminder that Watson relies on stored data, not true knowledge. The speed and percentage of correct answers should more than make up for doubts on the capabilities and importance of decision management and analytics shown with Watson, says James Taylor of the Smart Data Collective.
Outside of the game show, more than half of the show’s second of three broadcasts (a final, full match will go on today) was dedicated to IBM’s real-world expectations of Watson’s data prowess. Health care – a focus of IBM’s interests – was a primary focus. The database benefits and possibilities in that and other industries could be remarkable, says computer scientist Dr. Ira Haimowitz.
For one human – distant third-place competitor Ken Jennings – an immediate benefit would be a less robotic response to the game show’s buzzers. IBM developers plan to discuss those real-world applications and more through Twitter and a TED panel discussion Thursday morning.
Day Three – February 16
IBM’s supercomputer Watson became possibly the first nonhuman millionaire by besting its human competition on the final day of the Jeopardy challenge, and now is racking up interest in its analytics as well as critics doubting its functionality. Watson’s also gotten more than a few mentions as being the first sign of a terrifying robotic ruling class.
Watson excelled at times and seemed downright human at others (especially with the tricky answer-as-question wording that Jeopardy contestants must master) in the final day of the TV quiz show. All in all, the computer doubled the money totals of his two human competitors, along the way boosting IBM’s analytics efforts, donating money to disaster outreach programs with its winnings, not to mention pumping up ratings for the long-running game show.
Use in the health care industry cropped up as the show was wrapping up, with a collaboration planned with Nuance and two hospitals interested in the natural language data search displayed in Watson, and other medical leaders asking how they can harness it to quell very human health questions. Lawyers chimed in on ways Watson might parse legalese, and search engine expert Danny Sullivan wondered what the computer means for SEO, Google and Web data.
And, in his final Jeopardy response, second-place finisher (and human) Jennings welcomed a much farther-reaching application of Watson’s capabilities.