University researchers work to take computing to new levels
In a highly-publicized contest last year on “Jeopardy!,” IBM’s Watson computer came out as the winner against a series of highly-intelligent human champions. Despite Watson’s impressive showing against some of the most intelligent competitors in the history of the show, the computer is limited by its base model, a problem with computers that two professors at Missouri State University are working to change.
“Watson pushes the traditional Turing model of computation toward its limits, by using large computer processing, massive amounts of memory, enormous databases and clever programming,” said Dr. Emmett Redd, professor in the department of physics, astronomy and materials science at Missouri State University. “As powerful as Watson is, it is still incapable of updating its program to accomplish tasks other than those it is programmed for – it can’t adapt to a changing situation or learn how to do something new on its own.”
Redd and Dr. A. Steven Younger, research professor in the Center for Applied Science and Engineering in the Jordan Valley Innovation Center, are working on research to bring new capability and adaptability to modern computing. Together with Dr. Hava Siegelmann, associate professor at the University of Massachusetts – Amherst, Redd and Younger were awarded a $177,668 grant from the National Science Foundation (NSF) to research “Super-Turing Computation and Brain-Like Intelligence,” which could lead to more powerful, intelligent computers.
“Mathematician and World War II codebreaker Alan Turing developed a model of computation which has been the basis of computer design for the last seven decades,” said Younger. “Under that model, called the Turing Machine, enormous progress has been made, including IBM’s Watson, which is a direct descendant of Turing Machine-based computer. In the late 1990s, Dr. Siegelmann developed a theoretical model of a Super-Turing computation, which can perform computations that are beyond the abilities of a Turing Machine, including more brain-like intelligence.”
Redd and Younger have constructed a series of analog recurrent neural networks based on optical computing hardware, with their experimental results converging on the theoretical results of Siegelmann. The award from the NSF will allow them to bring together both sides of the research in a joint project with Siegelmann to develop a physical Super-Turing computer and evaluate its real-world computational capabilities.
“Where a Turing Machine is designed to do only what it is programmed for, Super-Turing is capable of trying different approaches, making mistakes and learning,” said Siegelmann. “This is what we are working on in our new research program.”
“Hot Topics – Expert Source” releases are a tool the office of university communications provides to assist media in locating a university source to comment on a particular subject or issue. The opinions expressed by the expert are those of a specific individual and are not necessarily representative of the views of the university.
High tech research has found a home at the Roy Blunt Jordan Valley Innovation Center (JVIC). The focus at JVIC, the anchoring facility for IDEA Commons, is environmentally friendly projects with an applied research emphasis on biomaterials, nanotechnology, carbon-based electronics, biomedical instrument development and energy.
New media resources
For more information about MSU, visit www.missouristate.edu. You can also be our fan on Facebook, follow us on Twitter, add us to your circle on Google+ and see what’s happening on YouTube. Receive notification of Missouri State news releases as soon as they are posted by subscribing to our RSS feed.