Neural Synaptic Weighting With a Pulse-Based Memristor Circuit

Hyongsuk Kim, Maheshwar Pd. Sah, Changju Yang, Tamás Roska, Leon O. Chua

IEEE Transactions on Circuits and Systems I Regular Papers · 2011 · 347 citations · 26 references

Concepts

TL;DR

A pulse‑based programmable memristor circuit is proposed to implement synaptic weights in artificial neural networks. The circuit uses five bridge‑connected memristors and charge‑dependent Ohm’s law to perform positive and negative multiplications, with shared input pulses for sign switching, weight setting, and synaptic processing, and is evaluated via simulations with linear and nonlinear memristor models.

Abstract

A pulse-based programmable memristor circuit for implementing synaptic weights for artificial neural networks is proposed. In the memristor weighting circuit, both positive and negative multiplications are performed via a charge-dependent Ohm's law (). The circuit is composed of five memristors with bridge-like connections and operates like an artificial synapse with pulse-based processing and adjustability. The sign switching pulses, weight setting pulses and synaptic processing pulses are applied through a shared input terminal. Simulations are done with both linear memristor and window-based nonlinear memristor models.

References

26