SPTRONICS

Adv. Mater., 2019, doi: 10.1002/adma.201900636

Neuromorphic computing is the development of computing schemes inspired by the processing of information in the brain, which can execute complex tasks very efficiently using an architecture that is completely different from that of semiconductor chips. Recently, researchers from Tohoku University have realized an artificial neuron and synapse in spintronics devices, which are promising for future energy-efficient and adoptive computing systems, as they behave like the spiking neural network in human brains.

The research group had previously developed a functional material system consisting of antiferromagnetic and ferromagnetic materials. This time, they prepared artificial neuronal and synaptic devices microfabricated from the material system, which demonstrated fundamental behavior of biological neuron and synapse, i.e., “leaky integrate-and-fire” and “spike-timing-dependent plasticity”, respectively, based on the dynamics of spin-orbit torque switching. The spintronics neuron and synapse could form a currently lacking hardware basis for neuromorphic computing, which would advance the fields of neuromorphic computation and neuroscience by enabling full-scale brain emulations, artificial intelligence, and human–machine interfaces.

Dahai Wei (Institute of Semiconductors, CAS, Beijing, China)

doi: 10.1088/1674-4926/40/5/050204