Researchers working on a computer that mimics the human brain
Researchers working on a computer that mimics the human
brain
Future computer systems could better tackle big data and
autonomous systems
By Sharon Gaudin
May 20, 2014 07:04 AM ET
Computerworld - Researchers at Sandia National
Laboratories are working on a computer that can tackle real-world situations in
real-time and can run on the same power as a 20-watt light bulb.
Right now the only "machine" that can handle
those functions is the human brain.
That's why scientists are trying to build a computer
system that works more like a brain than a conventional computer.
"Today's computers are wonderful at bookkeeping and
solving scientific problems often described by partial differential equations,
but they're horrible at just using common sense, seeing new patterns, dealing
with ambiguity and making smart decisions," said John Wagner, cognitive
sciences manager at Sandia National Laboratories, in a statement.
Scientists at Sandia, which are two major U.S. Department
of Energy research and development operations, are working on neuro-inspired
computing as part of a long-term research project on future computing systems.
"We're evaluating what the benefits would be of a
system like this and considering what types of devices and architectures would
be needed to enable it," said Sandia microsystems researcher Murat
Okandan. ""If you do conventional computing, you are doing exact
computations and exact computations only.
"If you're looking at neurocomputation, you are
looking at history, or memories in your sort of innate way of looking at them,
then making predictions on what's going to happen next," he added.
"That's a very different realm."
Neuro-computing systems are expected to be much
better-suited to taking on big data problems, which the U.S. government, along
with major enterprises, are working on. The systems also should be better at
handling remote autonomous and semiautonomous systems that need greater, and
different, computational power, as well as better energy efficiency.
Computers that function more like the human brain could
better operate unmanned drones, robots and remote sensors. They also are
expected to be ideal for handling complex analysis.
The systems would be able to detect patterns and
anomalies, sensing what fits and what doesn't, according to Okandan.
A neuro-inspired computer would vary in its basic
functioning from today's computer systems, which largely are calculating
machines with a central processing unit and memory that stores a program and
data. Today's machines take a command from the program and data from the memory
to execute the command, one step at a time. Of course, parallel and multicore
computers can do more than one thing at a time but still use the same basic
approach.
However, the architecture of neuro-inspired computers is
expected to be fundamentally different.
These future machines would be designed to unite
processing and storage in a single network architecture "so the pieces
that are processing the data are the same pieces that are storing the data, and
the data will be processed with all nodes functioning concurrently,"
Wagner said. "It won't be a serial step-by-step process. It'll be this
network processing everything all at the same time. So it will be very
efficient and very quick."
A neural-based computer architecture also would have far
more working connections.
Each neuron in a neural structure can have connections
coming in from about 10,000 neurons, Sandia Labs explained. However,
conventional computer transistors connect, on average, to four other
transistors in a static pattern.
Computer scientists have focused on mimicking the brain's
neural connections before but there's much more excitement about this project
because of the advances being made in the field.
Despite these advances, Sandia Labs noted that
researchers should be able to create the new architecture, in simple forms, in
the next few years. More complex systems may still be "decades" away.
Comments
Post a Comment