IBM experimental chips emulate the human brain
By Ben Coxworth
August 18, 2011
In April, the University of Southern California made the headlines when it announced that researchers there had created a functioning synthetic synapse circuit using carbon nanotubes. Well, today IBM unveiled a new class of experimental computer chips that are designed to emulate the human brain's abilities for perception, action and cognition. According to the company, "The technology could yield many orders of magnitude less power consumption and space than used in today's computers."
Utilizing advanced algorithms and silicon circuitry, the two prototype "neurosynaptic computing chips" are said to recreate the phenomena that takes place between spiking neurons and synapses in biological systems. The idea is that such chips would be used in "cognitive computers," which would learn through experiences - like the human brain - rather than simply being programmed.
To that end, IBM has joined forces with a number of academic partners, to develop such computers through the Systems of Neuromorphic Adaptive Plastic Scalable Electronics (SyNAPSE) project. According to the company, "The goal of SyNAPSE is to create a system that not only analyzes complex information from multiple sensory modalities at once, but also dynamically rewires itself as it interacts with its environment - all while rivaling the brain's compact size and low power usage." Phases 0 through 1 have already been completed, while the Defense Advanced Research Projects Agency (DARPA) has reportedly awarded the project US$21 million in funding for Phase 2.
The two chips themselves contain no biological components. According to the press release, however, both chips do feature 256 artificial neurons, with one core containing 262,144 programmable synapses, and the other containing 65,536 learning synapses. In lab tests, the chips have so far been used to execute simple applications such as navigation, machine vision, pattern recognition, associative memory and classification.
Ultimately, IBM hopes to produce a chip system featuring ten billion neurons and hundred trillion synapses, that would consume one kilowatt of power and have a volume of less than two liters (0.5 U.S. gallons).
"Future applications of computing will increasingly demand functionality that is not efficiently delivered by the traditional architecture," said Dharmendra Modha, project leader for IBM Research. "Imagine traffic lights that can integrate sights, sounds and smells and flag unsafe intersections before disaster happens or imagine cognitive co-processors that turn servers, laptops, tablets, and phones into machines that can interact better with their environments."
Partners in Phase 2 of SyNAPSE include Columbia University, Cornell University, the University of California at Merced, and the University of Wisconsin, Madison.