Self-Adaptive Spike-Time-Dependent Plasticity of Metal-Oxide Memristors

Abstract

Metal-oxide memristors have emerged as promising candidates for hardware implementation of artificial synapses – the key components of high-performance, analog neuromorphic networks - due to their excellent scaling prospects. Since some advanced cognitive tasks require spiking neuromorphic networks, which explicitly model individual neural pulses (“spikes”) in biological neural systems, it is crucial for memristive synapses to support the spike-time-dependent plasticity (STDP). A major challenge for the STDP implementation is that, in contrast to some simplistic models of the plasticity, the elementary change of a synaptic weight in an artificial hardware synapse depends not only on the pre-synaptic and post-synaptic signals, but also on the initial weight (memristor’s conductance) value. Here we experimentally demonstrate, for the first time, an STDP behavior that ensures self-adaptation of the average memristor conductance, making the plasticity stable, i.e. insensitive to the initial state of the devices. The experiments have been carried out with 200-nm Al2O3/TiO2−x memristors integrated into 12 × 12 crossbars. The experimentally observed self-adaptive STDP behavior has been complemented with numerical modeling of weight dynamics in a simple system with a leaky-integrate-and-fire neuron with a random spike-train input, using a compact model of memristor plasticity, fitted for quantitatively correct description of our memristors.

Memristor with Ag‐Cluster‐Doped TiO2 Films as Artificial Synapse for Neuroinspired Computing

“…Therefore, development of new memristor device is extremely necessary to improve learning and computing speed. [17,18,23,24] In this work, we propose that Ag nanoclusters are gradient assembled in TiO 2 films (named as TiO 2 :Ag composite films) by cosputterting at fabrication process. Using this composite films, we fabricate Ag/TiO 2 :Ag/Pt device and demonstrate the conduction can be gradually tuned under both of positive and negative pulse trains, respectively.…”

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

“…A dependency of the learning rules on the initial conductance was also presented for an Al2O3/TiO2-x memristor in Ref. 32. Finally, the present device allows controlling the time window for conductance modifications by tuning the device layout.…”

“…the Al2O3/TiO2-x memristors reported in Ref. 32. The steep current increase at Vd occurs due to a fast discharging mechanism, while below Vd the device operates in a slow discharging regime of the QDs.…”

“…Hence, the symmetric and asymmetric learning rules are beneficial for pattern completion and the recalling and storing of temporal sequences of action potentials, respectively. 28,29 The four different learning rules were artificially emulated by varying electrical input signals in chalcogenide 23,30,31 and metal oxide memristors 32 , and by varying optical input signals of metal-sulphide microfibers. 33 We present the emulation of four learning rules with a quantum dot memristor where the conductance change corresponds to charge transfer between quantum dots (QDs) and a two-dimensional electron gas (2-DEG).…”

Spike‐Timing‐Dependent Plasticity in Memristors

“…When G 0 is close to its maximum value, the increase of the conductance is very low. And when G 0 is close to its minimum value, the decrease of the conductance is very low [14]. …”