Researchers at Stanford University have created a computer board modeled after the human brain that is capable of simulating over 1 million neurons and billions of synaptic connections using the power of an iPad.
According to an article published on IEEE.org, they were able to achieve this by using 16 custom-designed “Neurocores” and placing them on a circuit board. The project is appropriately named Neurogrid; the circuit board resembles the human brain in the way it communicates with its circuits.
“If we can compute like the brain, then we can develop technology that has the power of the brain,” said Sam Fok, a grad student researcher on Neurogrid, in the video above.
Working in conjunction with Stanford’s Bio-X center, Neurogrid’s future applications could end up helping people who need artificial limbs or are paralyzed. Through implants in the brain, the processor would interpret movement data and translate the information to prosthetics.
Currently, these chips costs over $40,000 each. But Kwabena Boahen, a professor at Stanford leading the research, says that by switching over to bulk production the cost could drop to about $400.
Unfortunately we’ve still got a long way to go in matching the processing power and efficiency of the brain. The human brain has over 100 billion neurons running on 20 watts, the equivalent to a 60-volt light bulb. Neurogrid meanwhile runs one million neurons at about six watts.
As we see computer architecture take shape to match our own biological forms, it takes us a step closer to the Singularity, a term made popular by futurist Ray Kurzweil in which we inevitably integrate ourselves with robotics. We'll then be able to supersede our inherent capabilities or address the handicaps that disable us.
Julian Reyes is a VR Producer for Fusion.