Flexible three-dimensional artificial synapse networks with correlated learning and trainable memory capability

Chaoxing Wu,Dmitri B Strukov,Hwan Young Choi,Tae Whan Kim,J Joshua Yang

doi:10.1038/s41467-017-00803-1

Abstract

If a three-dimensional physical electronic system emulating synapse networks could be built, that would be a significant step toward neuromorphic computing. However, the fabrication complexity of complementary metal-oxide-semiconductor architectures impedes the achievement of three-dimensional interconnectivity, high-device density, or flexibility. Here we report flexible three-dimensional artificial chemical synapse networks, in which two-terminal memristive devices, namely, electronic synapses (e-synapses), are connected by vertically stacking crossbar electrodes. The e-synapses resemble the key features of biological synapses: unilateral connection, long-term potentiation/depression, a spike-timing-dependent plasticity learning rule, paired-pulse facilitation, and ultralow-power consumption. The three-dimensional artificial synapse networks enable a direct emulation of correlated learning and trainable memory capability with strong tolerances to input faults and variations, which shows the feasibility of using them in futuristic electronic devices and can provide a physical platform for the realization of smart memories and machine learning and for operation of the complex algorithms involving hierarchical neural networks.

Highlights

If a three-dimensional physical electronic system emulating synapse networks could be built, that would be a significant step toward neuromorphic computing
This means that the chemical synapses pass information directionally from a pre-synaptic cell to a post-synaptic cell, which is similar to the rectification behavior of a rectifier diode and provides a potential solution to suppress the crosstalk effect
A conceptual schematic diagram and a photograph of the flexible 3D-ASN based on copper-ion-doped pMSSQ (Cu-pMSSQ) are presented in Fig. 1a and Fig. 1b, respectively

Summary

Introduction

If a three-dimensional physical electronic system emulating synapse networks could be built, that would be a significant step toward neuromorphic computing. We report flexible three-dimensional artificial chemical synapse networks, in which two-terminal memristive devices, namely, electronic synapses (e-synapses), are connected by vertically stacking crossbar electrodes. The three-dimensional artificial synapse networks enable a direct emulation of correlated learning and trainable memory capability with strong tolerances to input faults and variations, which shows the feasibility of using them in futuristic electronic devices and can provide a physical platform for the realization of smart memories and machine learning and for operation of the complex algorithms involving hierarchical neural networks. The 3D interconnectivity, which plays a vital role in high-density information storage and multi-dimensional information processing in biological neural networks, has not yet been well realized with existing electronic devices[31]. The 3D-ASN is shown to be able to mimic correlated learning and exhibit a trainable memory function with a strong tolerance to input faults; this is accomplished by taking advantage of the programmable synaptic weights of the e-synapses

Methods

Results

Conclusion