Novel training method for metal-oxide memristive synapse device to overcome trade-off between linearity and dynamic range

Abstract

Synapse devices are essential for the hardware implementation of neuromorphic computing systems. However, it is difficult to realize ideal synapse devices because of issues such as nonlinear conductance change (linearity) and a small number of conductance states (dynamic range). In this study, the correlation between the linearity and dynamic range was investigated. Consequently, we found a trade-off relationship between the linearity and dynamic range and proposed a novel training method to overcome this trade-off.

Impact of Oxygen Reservoir Layer on 3T Oxygen Ion‐based Electrochemical Random Access Memory Performance

“…In addition, the linearity, defined as 1/(1+S), considering all conductance levels is approximately 0.8 for both potentiation (α p ) and depression (α d ). [ 22 ] S was derived by integrating all differences between ideal and measured conductance values (Figure , Supporting Information). The single layer exhibited relatively poor synaptic characteristics, strongly implying that gate leakage disturbance should be prevented to achieve reliable 3T Ox‐ECRAM synaptic characteristics.…”