SCIENTISTS AT IBM ZURICH and Swiss university ETH Zurich have come closer to building a machine that computes like a human brain.
The researchers found a way to imitate the actions of neurons, which fire electrical charges across a membrane where it is picked up by a synapse.
The human brain contains about 100 billion neurons together with many more neuroglia cells that support and protect them. In addition, each neuron may be connected to 10,000 other neurons with as many as a quadrillion synaptic connections.
The human brain is often described (by Earthlings, naturally) as the most powerful computer in the universe. It's capable, they point out, of performing complex calculations while drawing only about 20W of power and taking up a very modest two litres of head space.
Well, that may be so. But how come humans still lose their keys, fall down the stairs and forget the name of the person they were chatting up last night?
The new artificial neurons mimic the biological process of neurons firing signals at each other using common or garden materials including germanium antimony telluride, which are the basis of data storage on Blu-ray discs.
This material changes phase when an electrical pulse is applied, becoming crystalline and allowing the artificial neuron to fire.
The IBM team managed to bring a large number of these artificial neurons together and used them to perform simple computational tasks. The breakthrough, they claim, is down to the materials.
They are well understood, cheap, stable over billions of switching cycles, draw only a small amount of power and can be produced at the kind of nanoscales required to create microcircuits.
Like human neurons, the artificial ones function stochastically, meaning that there is a degree of randomness in their operation. It is hard to predict when a particular neuron will fire, but en masse they can be used to perform complex computations.
Large populations of these high-speed, low-energy nanoscale neurons could be used to create neuromorphic processors that are able to learn from their experience.
"Populations of stochastic phase-change neurons, combined with other nanoscale computational elements such as artificial synapses, could be a key enabler for the creation of a new generation of extremely dense neuromorphic computing systems," said scientist Tomas Tuma.
Well, that's all very well. But the real test will be how well they work after being soaked in beer. μ
This weeks in-brief Google News
To replace them with younger models
Security firm warns that IoT devices are the next target
But don't go expecting any new MacBooks