High‐Performance Neuromorphic Computing Based on Ferroelectric Synapses with Excellent Conductance Linearity and Symmetry

Abstract

Artificial synapses can boost neuromorphic computing to overcome the inherent limitations of von Neumann architecture. As a promising memristor candidate, ferroelectric tunnel junctions (FTJ) enable the authors to successfully emulate spike-timing-dependent synapses. However, the nonlinear and asymmetric synaptic weight update under repeated presynaptic stimulation hampers neuromorphic computing by favoring the runaway of synaptic weights during learning. Here, the authors demonstrate an FTJ whose conductivity varies linearly and symmetrically by judiciously combining ferroelectric domain switching and oxygen vacancy migration. The artificial neural network based on this FTJ-synapse achieves classification accuracy of 96.7% during supervised learning, which is the closest to the maximum theoretical value of 98% achieved to date. This artificial synapse also demonstrates stable unsupervised learning in a noisy environment for its well-balanced spike-timing-dependent plasticity response. The novel concept of controlling ionic migration in ferroelectric materials paves the way toward highly reliable and reproducible supervised and unsupervised learning strategies.

Synaptic and Gradual Conductance Switching Behaviors in CeO2/Nb–SrTiO3 Heterojunction Memristors for Electrocardiogram Signal Recognition

“…It is very important for the memristor to adjust the conductance continuously and evenly by positive and negative stimulation, which contributes to realizing the practical application. − As shown in Figure a, the positive and negative stimulation are induced by stepped pulses (Figure S10). The potentiation and depression processes are composed of 80 states and repeat for 100 cycles.…”

All-ferroelectric implementation of reservoir computing

“…Ferroelectric memristors are emerging memory devices which use the polarization switching to tune the resistance 28,29 . Because of the high controllability of polarization switching, ferroelectric memristors can exhibit highly reproducible memristive responses and potentially unlimited endurance [30][31][32][33][34] . Besides, they also show high switching speed and low power consumption [35][36][37] .…”

“…Besides, they also show high switching speed and low power consumption [35][36][37] . Owing to these advantages, ferroelectric memristors have recently been extensively investigated as synaptic devices for FNNs [31][32][33][34] . However, their use in RC systems is still scarce and mainly restricted to the reservoir 11,27,[38][39][40][41] .…”

Double-Floating-Gate van der Waals Transistor for High-Precision Synaptic Operations

“…14 Typically, linearity and symmetry of weight updates, which are realized via a highly precise programming process, offer training accuracy, network capacity, and discrimination among input signals in neuromorphic computing systems. 15 Two-dimensional (2D) materials are certainly promising candidates for post-CMOS technology owing to their extraordinary intrinsic properties, including the atomically thin structure and lack of surface dangling bonds, which ensure active external manipulations such as gating and minimized interface trapped charges. 16−23 Consequently, heterostructures composed of 2D materials have been widely proposed as promising candidates for artificial synapses.…”

“…This is due to the structural limitations of three-terminal transistors, which make it difficult to achieve linear output changes with identical input signals . Typically, linearity and symmetry of weight updates, which are realized via a highly precise programming process, offer training accuracy, network capacity, and discrimination among input signals in neuromorphic computing systems …”