Oct. 29 (UPI) -- Researchers have created a new computer component capable of toggling between 16 possible memory states -- the kind of computing versatility provided by brain synapses.
The new component, called a graphene field effect transistor, described Thursday in the journal Nature Communications, could pave the way for advances in brain-inspired computing.
Modern computers are exclusively digital, featuring two states: on-off or zero and one. Engineers at Penn State University are working to build a computer that replicates the brain's analog nature, capable of hosting many different states.
If a digital computer's information processing components work like a light switch, toggling only between on and off, then an analog computer is like a light dimmer.
Scientists have been investigating the potential of brain-based computing for decades, but analog computers have been overshadowed by the advances in traditional computing power. However, the rise of big data and smart devices like self-driving cars has highlighted the need for more computing efficiency.
"We have powerful computers, no doubt about that, the problem is you have to store the memory in one place and do the computing somewhere else," lead researcher Saptarshi Das, an assistant professor of engineering science and mechanics at Penn State, said in a news release.
All the movement of information required by the bifurcation of memory to logic in modern computers puts a strain on speed. It also requires more spaces. Das and his research partners estimate that their graphene field effect transistor can help eliminate this bottleneck.
"We are creating artificial neural networks, which seek to emulate the energy and area efficiencies of the brain," said study first author Thomas Shranghamer.
"The brain is so compact it can fit on top of your shoulders, whereas a modern supercomputer takes up a space the size of two or three tennis courts," said Shranghamer, a doctoral student in the Das group.
Brain synapses can be quickly reconfigured to create a variety of neural network patterns. Likewise, the new graphene field effect transistor, formed by a one-atomic-thick layer of carbon atoms, can be used to control 16 possible memory states.
Researchers were able to reconfigure the transistor, effectively toggling between memory states, by applying a brief electric field to the graphene layer.
"What we have shown is that we can control a large number of memory states with precision using simple graphene field effect transistors," Das said.
Das and his research partners are now looking to work with semiconductor companies to attempt to scale-up the production of the new technology.
