Mimicking of pulse shape-dependent learning rules with a quantum dot memristor

Abstract

We present the realization of four different learning rules with a quantum dot memristor by tuning the shape, the magnitude, the polarity and the timing of voltage pulses. The memristor displays a large maximum to minimum conductance ratio of about 57000 at zero bias voltage. The high and low conductances correspond to different amounts of electrons localized in quantum dots, which can be successively raised or lowered by the timing and shapes of incoming voltage pulses. Modifications of the pulse shapes allow altering the conductance change in dependence on the time difference. Hence, we are able to mimic different learning processes in neural networks with a single device. In addition, the device performance under pulsed excitation is emulated combining the Landauer-Büttiker formalism with a dynamic model for the quantum dot charging, which allows explaining the whole spectrum of learning responses in terms of structural parameters that can be adjusted during fabrication such as gating efficiencies and tunneling rates. The presented memristor may pave the way for future artificial synapses with a stimulus-dependent capability of learning.

Brain‐Inspired Structural Plasticity through Reweighting and Rewiring in Multi‐Terminal Self‐Organizing Memristive Nanowire Networks

“…Previous attempts showed that such synaptic interactions can be emulated in top‐down fabricated devices based on Ag nanoclusters or 2D materials . Also, modulation of synaptic activity was reported by pulse shape engineering or light stimulation . Here, Ag‐NW network shows that the stimulation of a synaptic pathway (S10, Supporting Information) results in synaptic weight changes not only in the directly stimulated synapse but also in other nonstimulated synaptic pathways (Figure c), thus mimicking heterosynaptic facilitation.…”

Semiconductor Quantum Dots for Memories and Neuromorphic Computing Systems

“…Hartmann et al demonstrated this concept via a sitecontrolled QD memory, in which four types of learning rules (asymmetric Hebbian, asymmetric anti-Hebbian, symmetric Hebbian, and symmetric anti-Hebbian) were realized. 420 The device consisted of a channel, a lateral four-gate system, and site-controlled QDs above the channel. Figure 29f depicted the device structure and the circuit diagram, and V pr and V po pulses were employed to mimic the pre-and postsynaptic input signals, respectively.…”

“…(g) The emulation of learning rules by the QD memristor achieved by voltage pulses with different shapes, amplitudes, and timing. Reproduced with permission from ref . Copyright 2016 AIP Publishing.…”

“…Research efforts have been performed to endow artificial synapses with stimulus-dependent capability, thus enabling us to emulate different types of learning rules in a single memory device. Hartmann et al demonstrated this concept via a site-controlled QD memory, in which four types of learning rules (asymmetric Hebbian, asymmetric anti-Hebbian, symmetric Hebbian, and symmetric anti-Hebbian) were realized . The device consisted of a channel, a lateral four-gate system, and site-controlled QDs above the channel.…”

Hysteresis Control in Hybrid Perovskite-based Devices for Photovoltaic and Neuromorphic Computing Applications

“…It has a compositional space of >10 6 formulations that can be further explored via a facile solution-based processing [25][26][27][28]. Moreover, the integration of low-dimensional structures [110][111][112], mixed formulations [113,114], nanocrystals [22,115] and quantum dots [116,117] further expands the already vast degrees of freedom or internal state variables in perovskite-based memristors. In addition, the electrochemical reactivity of the mobile ion species with the top and bottom electrodes proves to be crucial in the switching type and performance of perovskite-based memristors [118][119][120].…”