Tunable flexible artificial synapses: a new path toward a wearable electronic system

Abstract

Artificial synapses: memristive transistors with mechanical and synaptic flexibility Mechanically flexible artificial synapses based on memristive transistors demonstrate different kinds of synaptic plasticity. The synapse is a fundamental component in neuromorphic computing (a brain-inspired computing approach that aims to provide more efficient computing compared to conventional approaches). Yiqiang Zhan, Lirong Zheng, and Fernando Seoane with collaborators in Sweden and China now report an artificial synapse based on a memristive transistor that is mechanically flexible. Key to the design of their synapse is a three-terminal structure, which enables gate tuning. The ability to adjust the voltage on the gate terminal enables variations in the device to be compensated thereby improving performance uniformity and repeatability. The researchers also show that gate tuning can reduce the total energy consumption per spiking event to 45 fJ and demonstrate a variety of synaptic plastic features important for replicating neuromorphic behavior.