Nanoscale device powers synthetic brain

University of Southampton scientists have demonstrated that a nanoscale device could power artificial systems to mimic a human brain.

Memristors, first theorised in the 1970s, are electrical components that limit or regulate the flow of electrical current in a circuit. They can ‘remember’ the amount of charge that flows through it and even retain data when the power is turned off.

In this study, published in Nature Communications, they formed part of an artificial neural network (ANN) - systems capable of exhibiting learning abilities and performing tasks difficult for conventional computing systems. The researchers used memristor synapses in an ANN to carry out reversible learning of random data.

Dr Alex Serb, lead author and from the University of Southampton, said: “If we want to build artificial systems that can mimic the brain in function and power we need to use hundreds of billions, perhaps even trillions of artificial synapses, many of which must be able to implement learning rules of varying degrees of complexity. If artificial brains are ever going to become reality, therefore, memristive synapses have to succeed."

The memristor array was capable of learning and re-learning input patterns in an unsupervised manner within a probabilistic winner-takes-all network — commonly used in computational brain models. This can allow for low-powered processors to process big data in real-time.

Dr Themis Prodromakis, reader in Nanoelectronics, also from the University of Southampton, said: “Obvious applications for memristors are emerging memory technologies. Other alternatives include their use as dynamic resistive loads in tunable circuits.”