Last modified: Wednesday, June 30, 2010
IU natural computing guru Mills gets chapter treatment in exploration of computing's outer reaches
FOR IMMEDIATE RELEASE
June 30, 2010
BLOOMINGTON, Ind. -- Jonathan Mills is not your average computer scientist. When new computers start rolling into Indiana University Bloomington's School of Informatics and Computing where he is an associate professor, Mills' mouth starts to water, but not over motherboards, CPUs or RAM.
Instead Mills, one of 15 computer scientists highlighted in a new book called Natural Computing, tosses aside the power supplies and external devices. The hybrid computer designer -- the only researcher spotlighted in the book who is not based on either the East or West coast -- has his eyes on packing foam inside the box that could soon become working parts in his next machine.
Since 1990, Mills has been studying naturally-occurring information processes, most specifically in the form of extended analog computers (EACs) that, rather than decompose data into pieces called bits, use a family of devices to compute specific functions by analogy.
That family of devices can include components as varied as foam, soap film, Plexiglas, and even Jell-O, to perform physical computations. The research is founded in a diverse range of studies that include computational complexity, very-large-scale integration design, neurobiology, chaos and non-linear dynamics, information theory, mathematical biology, quantum mechanics and statistical thermodynamics.
"If this seems like a different way to think about computing, Jonathan Mills is a different kind of computer scientist," write Natural Computing authors Dennis Shasha and Cathy Lazere.
Mills' work focuses on the design, applications and theory of extended analog computers developed by the late mathematician Lee A. Rubel, whose interest in brain function led him to theorize that just like the brain models space and time through solving the equivalent of partial differential equations, a machine might too.
Rubel thought such a machine could never be built, but Mills and Bryce Himebaugh, a computer design engineer at IU, have built a collection of them, including a recent one that delivers electrical current to one of 25 points on a conductive sheet of foam with a five-by-five input and output array of LEDs and programmable electrical current sources and sinks. With the foam serving as the primary computing element, current is taken from specific points, or current sinks, and the voltage measured or fed to an output circuit.
The device also contains analog-to-digital and digital-to-analog converters for hybrid computing, a USB interface to a host digital computer and a microprocessor that controls the five-by-five array with stored configuration data (but not a stored memory program where an algorithm usually performs computations).
"It is a paradigm that complements digital computing, and unifies digital and analog computer architectures in a very satisfying way," Mills said.
Taking problem-solving scenarios from everyday challenges like weather modeling, smarter investment strategies and new medical techniques, the authors use Natural Computing to guide readers through a series of methodologies on biological computing that come straight from Mills and other researchers.
"If you want a device that will repair skin, bones or arteries it makes far more sense to build the device out of DNA, viruses or cells than to build it out of electronics," Shasha and Lazere said in announcing release of the book published by W. W. Norton & Co.
Mills goes so far in the book as to explain through a simple kitchen experiment using baking soda, vinegar, food coloring, paper towels, cups and a pan, how an EAC can compute butterfly wing morphogenesis. It's a simple experiment that models chemical reaction-diffusion equations.
"Suppose you have a pre-teen or teen or even someone at a high school or community college who is interested in this topic. There is only one chapter in the book that describes a device the kid can physically build at home and get a rough idea of how it works," Mills said. "So I realized that I am writing to the future scientists, not to my peers, and of all the people in the book I have the best chance at reaching young minds."
About the authors of Natural Computing: Shasha is a professor of computer science at the Courant Institute of New York University and writes Scientific American's "Puzzling Adventures" column. Lazere is a writer and former editor at the Economist Intelligence Unit.
To speak with Mills, Shasha or Lazere, please contact Steve Chaplin, University Communications, at 812-856-1896 or firstname.lastname@example.org.