By modeling a circuit control panel on the human brain , Stanford bioengineers have developed microchips that are 9,000 time faster than a typical personal computer . Called Neurogrid , these energy - efficient circuit could eventually power sovereign robots and innovative prosthetic limbs .
Bioengineers are smart to take inspiration from the human learning ability . It ’s a extremely effective information processor capable ofcrunching 100 million instruction per second(MIPS ) . Astoundingly , it only habituate about 20 watts to power its 100 billion neurons . Today , our estimable supercomputer require a million watts to imitate a million neuron in real time ( measured in terraflops ) . A standard desktop computing equipment requiresabout 40,000 clip more power to run and operates about 9,000 multiplication slower .
The destination , therefore , is to acquire information technologies with the magnate of the human brain . There are several initiatives underway that are working to achieve this destination , includingIBM ’s neurosynaptic chips(andaccompanying programming language ) , the University of Heidelberg’sHICANN Chip , and Einstein - function initiatives like the EuropeanHuman Brain Project .
We can now add another labor to the list : Stanford ’s Neurogrid . But unlike other current efforts , this “ neuromorphic ” system shoot a line some incredible energy - salve feature .
An Analog State of Mind
The new circuit board , developed by Kwabena Boahen and his workfellow at Stanford , consist of 16 custom - design “ Neurocore ” chip working in a tree web configuration . Each of the 16 Neurocores supports 65,536 neurons . Together , these chips can simulate one million neurons and 1000000000 of synaptic connexion . And fantastically , Neurocore needs just three James Watt of power to get the line done .
The designers used traditional transistors , but or else of using digital logic , they used them as analog circuit . To mime the functions of the human head ( albeit on a drastically reduced scale leaf ) , the research worker emulated all neural elements ( except the chassis ) with shared electronic circuits — a design decision that maximized the number of synaptic connections . To maximise vigour efficiency , the researchers used analog tour . And to maximise throughput , they complect the neural arrays in a Sir Herbert Beerbohm Tree web .
It ’s considered the most cost - in effect way to simulate neuron . But at $ 40,000 a piece , the researcher are going to have to figure out a way to repulse the costs down .
Miniaturization, Autonomy, Power
Ramped - up and processed versions of this engineering science could be put to good use . In addition to meliorate our understanding of how the human brain works , it could be used to see signal from the brain and , in genuine time , use those signals to drive prosthetic limbs for paralyzed people .
These chips could also be used in robotics . A automaton implanted with a Neurocore - like bit would n’t have to be tether to a power supply , thus increase its self-reliance .
Read the entire study at Proceedings of the IEEE : “ Neurogrid : A Mixed - Analog - Digital Multichip System for Large - Scale Neural Simulations . ” supplementary entropy viaStanford .
Cognitive scienceFuturism
Daily Newsletter
Get the best technical school , scientific discipline , and culture news program in your inbox daily .
tidings from the future , rescue to your present .
You May Also Like
