New Neural Chip Mimics Brain Function

November 19, 2012
By George I. Seffers
E-mail About the Author
Ronald E. Meyers delves into quantum physics research at the U. S. Army Research Laboratory. Meyers, fellow researcher Keith Deacon and Gert Cauwenberghs, a professor of bioengineering and biology at the University of California at San Diego, earned a patent for a futuristic neural computer chip.

Researchers working for the U.S. Army have developed and patented a neural computer chip that mimics human brain functions and could potentially be used for quantum computing, which could harness the power of individual atoms to perform functions millions of times faster than conventional computers. The new technology could eventually prove valuable for encryption and decryption and a wide range of other uses, including a next-generation Internet and the possibility of helping wounded soldiers better control prosthetic limbs.

Two Army Research Laboratory (ARL) scientists—Keith Deacon and Ronald Meyers—patented the neural chip for the Army in September. They worked with Gert Cauwenberghs, a professor of bioengineering and biology and co-director of the Institute for Neural Computation at the University of California at San Diego.

The technology uses synaptic connections for interfacing neurons and learning through feedback. “Some people would call it artificial intelligence, but it’s really a cooperative. The neurons are working together and cooperating with each other to try to learn and accomplish some task in an efficient manner,” explains Meyers, who also serves as lead investigator on the project. “It’s not yet quantum. It’s just not traditional.”

Meyers adds that the technology has a wide variety of potential uses once further developed. “Some of the applications might include eventually developing a faster, quicker, smarter Internet. And we also look for potential applications in assisting other biological systems. For example, if we have injured soldiers or other people with disabilities, perhaps some of the computing in this type of processor can assist in some tasks. One might think of prosthetics control or something related to the senses. It would take us working with others to make some of these applications come to fruition,” he says.

The real breakthrough with the neural chip is the mathematics used. “It’s a little different from other chips. It’s analogue, but it uses a different type of mathematics that control the processors. So, the chip has the capability, if expanded, to take in information and make quick decisions,” Meyers explains.

Conventional computer chips rely on mathematics named after German mathematician Rudolf Lipschitz, who died in 1903. The brain, however, uses non-Lipschitz processes capable of rapidly changing functions, which conventional chips simply have not handled well. “Another chip might go unstable if it tried to handle these types of functions,” Meyers indicates. “This chip was made with classic or conventional technologies, but the mathematics were nonconventional. We’re looking to move it to quantum, which of course, is nonclassical and quite different from the type of math we’re used to in our everyday lives,” Meyers states.

The ARL, like other military laboratories, engages in early, basic research. Once technology seems feasible and practical, it is transitioned to other organizations, which means it is largely up to the other organizations to find and develop the technology for specific uses.

Meyers declines to categorize the neural chip as either a small step or a major leap toward quantum computing. “I think we’ll determine that after we see the progress. I would just say that it looks very promising.”

Share Your Thoughts: