Brain-Inspired Molecular Memristors Could Boost the Efficiency and Performance of Future Computers

A team of physicists have published a paper detailing a new type of memory device, inspired by the human brain, which they believe could both simplify semiconductor design and offer future computers a big boost in both performance and efficiency.

"This work is a significant breakthrough in our quest to design low-energy computing," claims project lead Ariando, associate professor at the National University of Singapore's Department of Physics, of the work. "The idea of using multiple switching in a single element draws inspiration from how the brain works and fundamentally reimagines the design strategy of a logic circuit."

What the team has developed is a molecular memristor, an electronic device named as a combination of "memory" and "resistor" and used as the basis for next-generation computer memory, which they say shows reconfigurability far beyond its rivals — proven in a prototype which sandwiches a 40nm-thick layer of molecular film between gold and gold-infused nanodisc and indium tin oxide.

"Similar to the flexibility and adaptability of connections in the human brain, our memory device can be reconfigured on the fly for different computational tasks by simply changing applied voltages," explains first author Sreetosh Goswami, PhD, of the prototype. "Furthermore, like how nerve cells can store memories, the same device can also retain information for future retrieval and processing."

"This new discovery can contribute to developments in edge computing as a sophisticated in-memory computing approach to overcome the von Neumann bottleneck," Ariando claims, "a delay in computational processing seen in many digital technologies due to the physical separation of memory storage from a device's processor."

To prove that their creation could be used as a functional memory, the team had it run a range of real-world workloads — and showed that it could run complex computations in a single step then be immediately reprogrammed for a different task, in theory replacing thousands of transistors in a general-purpose processor with a single molecular memory device.

The team's work has been published in the journal Nature, under closed access terms.