A possible glimpse at a future form of high performance edge computing – networks of artificial brain synapses – developed by engineers at the Massachusetts Institute of Technology is showing promise as a new memristor design for neuromorphic devices, which mimic the neural architecture in the human brain.
Published today in Nature Nanotechnology, results of the engineers’ work show the potential for brain-inspired circuits to be built into small devices and execute complex tasks that today require a supercomputer, according to MIT.
The “brain-on-a-chip” is smaller than a piece of confetti and is comprised of tens of thousands of artificial synapses – known as memristors, short for memory transistors.
“So far, artificial synapse networks exist as software. We’re trying to build real neural network hardware for portable artificial intelligence systems,” said Jeehwan Kim, associate professor of mechanical engineering at MIT in a blog published today on the MIT News Office site. “Imagine connecting a neuromorphic device to a camera on your car, and having it recognize lights and objects and make a decision immediately, without having to connect to the internet. We hope to use energy-efficient memristors to do those tasks on-site, in real-time.”
Funded by the MIT Research Support Committee funds, the MIT-IBM Watson AI Lab, Samsung Global Research Laboratory and the National Science Foundation, among other sources, a key finding of the research is the use metallurgy techniques to develop an alloy made up of silver, copper and silicon that stabilizes ions used to transmit information from one memristor to another.
While current memristor designs function well when voltage stimulates a heavy flow of ions from one electrode to the other, they don’t perform as well when memristors generate subtler signals, according to MIT. “The…lighter the flow of ions from one electrode to the other, the harder it is for individual ions to stay together. Instead, they tend to wander from the group, disbanding within the medium. As a result, it’s difficult for the receiving electrode to reliably capture the same number of ions, and therefore transmit the same signal, when stimulated with a certain low range of current.”
Enter metallurgy, the science of melding metals and studying the properties of the resulting alloys.
“Traditionally, metallurgists try to add different atoms into a bulk matrix to strengthen materials, and we thought, why not tweak the atomic interactions in our memristor, and add some alloying element to control the movement of ions in our medium,” Kim says. The team found copper to be the right alloying element, as it is able to bind both with silver, and with silicon. “It acts as a sort of bridge, and stabilizes the silver-silicon interface,” Kim says.
All of which could enable the memristor to “work along a gradient, much like a synapse in the brain. The signal it produces would vary depending on the strength of the signal that it receives. This would enable a single memristor to have many values, and therefore carry out a far wider range of operations than binary transistors,” MIT reported. “Like a brain synapse, a memristor would also be able to ‘remember’ the value associated with a given current strength, and produce the exact same signal the next time it receives a similar current. This could ensure that the answer to a complex equation, or the visual classification of an object, is reliable — a feat that normally involves multiple transistors and capacitors.”
For further details, go to Jennifer Chu’s article at the MIT News Office.
