Beyond Jeopardy: putting Watson to work is a hit with students
REBECCA GUERRERO | January 25, 2012 | UNIVERSITY COMMUNICATIONS
He is part of a team that contributed to the creation of Watson, a supercomputer that astonished the world when it competed on and won the popular game show Jeopardy and he was also the latest speaker in UTEP's Centennial Lecture Series, Beyond Jeopardy. Richard Talbot, director of Product Line Management of IBM Power Systems, spoke to UTEP faculty, staff and students about how Watson came to be and the impact its creation could have on the future of technology.
"Watson is not a fantasy prop in a science fiction movie. It is a real computer with artificial intelligence capable of responding to Jeopardy questions posed not in binary code, but in natural human language with all of its complications," UTEP President Diana Natalicio said. "Watson is not only capable of answering such questions, it's really good at it."
Watson, named for IBM's founder Thomas J. Watson, was pitted against two of the most successful contestants in Jeopardy history last Feburary, including Jeopardy legend Ken Jennings. The result left many astounded. Watson not only won, it won by a large margin. Furthermore, Watson did not cheat by getting help through any sort of Internet access; it relied entirely on its own capacity.
"Watson represents an accomplishment that at one point in time we thought impossible to achieve," Talbot said. "The challenge was to retrieve, analyze and retain massive amounts of information, which here means about a million books, and then come back with accurate answers very quickly."
Talbot said Watson's capabilities are very different from the Google search engine, because when people use Google they typically receive many answers that need to be sort through to find the information required. Watson responds to questions quickly and with a high level of confidence that it is the correct answer.
According to Talbot, another characteristic that sets Watson apart is its ability to learn from its mistakes. While on the show Jeopardy, Watson got a few of the questions wrong because of the way Watson phrased the answers. After its feedback let it know that it was guessing incorrectly, it used a system of algorithms to change its strategy and went on to win the game without missing more questions, a feat that has even technical-minded people baffled.
"I'm very interested in this subject matter, I'm even taking a class on artificial intelligence this semester," said Jaime Daniel Peña, junior computer sciences major. "This is the future."
After Watson's success on Jeopardy, IBM has challenged itself to dedicate Watson's capabilities to solving real-world issues beginning with issues in health care.
"You can just imagine the possibilities for this type of technology," Talbot said. "You could give Watson two or three pieces of information regarding a patient's symptoms and it could respond with a possible diagnosis. At the same time it could tell you 'I'm not very confident with this answer. If I had two more symptoms I could provide you with a more confident prognosis', This would create a real life situation where there is more dialogue between the system and the specialist, and they are more likely to come up with a correct answer."
Talbot said Watson's technology is going to revolutionize the health care industry in a few ways. First, many cancer patients are referred to larger treatment centers only to realize they've been treated for the wrong form of cancer or disease processes in the past. Watson would help to eliminate such confusion with direct and accurate prognosis. Second, Watson would give the patients themselves more control over their own treatment.
"Watson can help patients to ask the right questions about their own symptoms. They will become smarter consumers about their practices and they can make sure to find the best sources of care," Talbot said.
During the Q&A session, some of the questions dealt with people's concern about Watson's artificial intelligence and many were worried that a supercomputer that is able to understand unstructured human language and learn from its mistakes. They said they might also be able to expand its knowledge base and eventually become more intelligent than human-kind, throwing us into an apocalyptic situation straight from the big screen.
Talbot, however, was quick to alleviate people's concerns by stressing that Watson was made for purely helpful purposes and as an informational resource.
"The type of technology it would take for a machine to be able to truly think for itself is something I'm not even familiar with," Talbot said.
Many of the students at the event came to see the super-computer that made headlines after its performance on Jeopardy and to learn what was behind the machine.
"I watched the jeopardy game, so I really wanted to find out what was behind it," said Hoong Yan See Tao, graduate electrical and computer engineering student. "I think the most exciting part is the things it can do in the future. I mean, they already sold the system to some industries, so the possibilities of how it can actually help human-kind are endless."
Every student, whether they attended for class credit or because of a personal interest, were able to take something away from the presentation.
"What I really found the most interesting was that this supercomputer can go back and use algorithms to correct itself. When you think about it, that is really something huge because in structured environments machines aren't able to do that," said Kathleen Zurlinden, graduate systems engineering student.
Talbot was confident that this kind of technology is going to solve meaningful and significant problems in today's society, and that in a matter of 10 to 20 years, many people will already have benefited from it.
"You'll be using this technology in the near future, and your children will benefit from it as well," Talbot said.