‘Brain’ on a microchip getting closer with TrueNorth breakthrough
Breakthrough in computer technology offers massive precessing power with none of the enormous power needs of the supercomputers
Computer science is edging close to manufacturing a brain, or at least an electronic cognition machine that operates as closely as possible to the speed and efficiency of the human cortex.
A coalition of IBM's research institutes and several universities and government laboratories delivered a preliminary answer on Thursday with a 5.4-billion transistor chip with 1 million programmable neurons and 256 million synapses.
The TrueNorth chip is the size of a postage stamp and 1,000 times as energy efficient as a conventional chip, according to a study published online on Thursday in the journal Science.
Don't expect to see the tiny supercomputer on your smartphone in the near future, although the lead researcher said his team was gaining momentum in that direction. He envisions a world populated with sensors that can process data at brain-like speeds, serving as guides for the blind or instant detectors of industrial toxins.
Modha and many others have been battling for decades because, for all the advances in processing speed, materials and manufacturing, digital computing relies on architecture from the 1940s. It has a well-known "bottleneck" between the processor and memory, named after the architect himself, John von Neumann.
Supercomputers that have hurdled the von Neumann bottleneck have accomplished stunning feats, but also have energy requirements that vie with some towns, and have grown larger than the laboratory-sized calculating machines of the infancy of computers.
"We have instrumented the planet with cameras, microphones, smartphones, a variety of sensors and the data is coming at us fast and furious," Modha said. "Asking today's computers to understand this sensory tsunami is architecturally very, very expensive."
Efforts to mimic the human brain on silicon have been underway for several years. This one, dubbed SyNAPSE, received US$53.5 million from the US Defence Advanced Research Projects Agency (DARPA). In 2011, it unveiled a miniscule "core" of 256 neurons and 262,000 synapses.
TrueNorth builds those cores into a system that can be expanded virtually without limits, mimicking the way interlaced neurons relay information via "spikes" in activity.
Like the brain, TrueNorth is event-driven. It conserves energy by doing only what is necessary to the task at the time, and no more, unlike conventional processors. "It doesn't have to run all the time," Modha said. "It's very parsimonious, like nature."
Modha, who spoke by phone, said he was holding a prototype chip in his hand, and his excitement was such that he dropped it. He seemed unperturbed.
Then again, he was so confident of the team's design, he put its manufacture on a one-month hiatus and promised a US$1,000 bottle of champagne to any member who could detect a flaw in the design. "No one claimed it. The chip came back from fabrication and worked flawlessly," he said
Researchers tested the chip by running a program to detect and identify people and vehicles while they move in a complicated environment at the Stanford University campus. It passed, and did so while consuming fractions of the energy of a supercomputer.
Modha said his team was not aiming to create an artificial brain, nor to replace von Neumann computers. One was impossible, the other impractical, he said. And the ability to learn by adapting and changing structure and function remains a purely cerebral talent that has not been fully tested in the new chip. Still, Modha said that capacity "is on our horizon".