Scientists at an IBM-led collaboration have produced a new computer chip dubbed ‘TrueNorth’ that mimics the organization of the brain, and squeezes in one million computational units called "neurons". The journal Science, which first published this feat, described it as a supercomputer the size of a postage stamp.
Describing the chip as “a new machine for a new era,” Dr. Dharmendra Modha, the senior author of the publication, said that this was the “cumulative total of over 200 person-years of work.”
The building blocks for the TrueNorth chip are ‘neurosynaptic cores’ of 256 neurons each, which IBM launched in 2011. What Dr. Modha and his team have managed to do is engineer an interconnected 64-by-64 grid of these cores on to a single chip, delivering over one million neurons in total. Because each neuron is connected to 256 others, there are more than 256 million connections or "synapses".
Unlike traditional chips that power nearly all modern computers and have been around since the early 1940s, TrueNorth will take some time to be commercially useful, because programs have to be written from scratch to run on this type of chip.
Current chips, which use binary ones and zeros as units of computation and have a design where the processors and memory are separate, are a natural match for sequential, mathematical operations. The new ‘neuron’ chips use ‘spikes’ as computational units and have a highly interconnected design that makes them much more efficient at handling lot of data at the same time.
The ‘spikes’, generated on neurons when its inputs are active enough, are sent across the chip to other ‘neurons’ taking them closer to their own threshold. Software has to be written completely differently for these spiking-network systems.
The complexity of ‘TrueNorth’ is impressive for a man-made device just 3cm across, but still pales in comparison with the human brain it attempts to emulate. Biological neurons, packed inside the brain, send and receive something in the order of 10,000 connections each.
But Dr. Modha is quick to point out that the new chip is "endlessly scalable". Multiple units can be plugged together to form another, still more powerful assembly. "This isn't a 10-15 percent improvement," he said. "You're talking about orders and orders of magnitude." Dr. Modha envisages myriad next-generation applications, from glasses that help visually impaired people navigate, to robots for scouring the scene of a disaster.