OK Watson, Game's Over. Now, Get to Work
By Susan Nunziata Watching IBM's supercomputer Watson take on Jeopardy! champs Ken Jennings and Brad Rutter these past three days has been my own personal, geeky version of the Super Bowl (minus a painful rendition of the National Anthem and an over-the-top halftime show).
It's not that I've been necessarily rooting for either man or machine. Instead, I've been marveling at the incredible abilities of the human mind, as well as the incredible feat of human engineering that is Watson. OK, so it's true: I'm a Watson groupie. Talk nerdy to me.
At an IBM-hosted Jeopardy! viewing party in New York City last night, Dr. David Ferrucci, the scientist leading the IBM Research team that has created Watson, shared some of amusing and unexpected challenges in prepping the computer for its big showdown.
For example, Watson was fed the equivalent of roughly 200 million pages of content from sources such as encyclopedia, dictionaries, bibles, literary classics, and news articles. Most of that information was ingested in text that is uppercase/lowercase, just like this article. Somewhere along the line, Ferrucci said, the team was informed that the Jeopardy! clues would, in fact, be fed to the machine in ALL CAPS, just as they appear on the TV screen. Ferrucci's response? "Crap!" He added: "That matters a lot." The Watson crew enlisted IBM colleagues who work on its "True Case" application to come up with a solution before big game.
Deciding which information to keep and which information to jettison was also part of the preparation process, said Ferrucci. "We toyed with the 'Urban Dictionary,' " he said. "But that caused us problems because of the pornographic content."
Watson is powered by 10 racks of IBM Power 750 servers running Linux. It uses 15 terabytes of RAM, 2,880 processor cores and is capable of operating at 80 teraflops. IBM clearly didn't put years of research dollars and brainpower into Watson simply to snag a killer product placement on a popular gameshow.
In fact, the company today kicked off a joint research project with Nuance Communication to explore, develop and commercialize the Watson computing system's advanced analytics capabilities in the healthcare industry. The research and technology initiative will combine IBM's Deep Question Answering (QA), Natural Language Processing, and Machine Learning capabilities with Nuance's speech recognition and Clinical Language Understanding (CLU) solutions for the diagnosis and treatment of patients. The two companies expect the first commercial offerings from the collaboration to be available in 18-24 months.
Columbia University Medical Center and the University of Maryland School of Medicine are also participating in the research project. According to IBM, physicians at Columbia University are helping identify critical issues in the practice of medicine where the Watson technology may be able to contribute. Physicians at the University of Maryland are working to identify the best way that a technology like Watson could interact with medical practitioners to provide maximum assistance.
Ferrucci said that in real applications, such as healthcare, Watson's value comes in offering up a complete "evidence profile" with every answer. This would give the user insight into what evidence Watson used and how it weighted that evidence to come up with a given conclusion.
"Human intelligence is a whole other league," says Ferrucci. "We don't want computers making value judgments about what it means to be human. Only humans can do that."
We couldn't agree more.