Adaptive Properties of Spiking Neuromorphic Networks with Synapses Based on Memristive Elements

Neuromorphic computing networks (NCNs) with synapses based on memristors (resistors with memory) can provide a much more effective approach to device implementation of various network algorithms as compared to that using traditional elements based on complementary technologies. Effective NCN implementation requires that the memristor resistance can be changed according to local rules (e.g., spike-timing-dependent plasticity (STDP)). We have studied the possibility of this local learning according to STDP rules in memristors based on (Co0.4Fe0.4B0.2)x (LiNbO3)1 –x composite. This possibility is demonstrated on the example of NCN comprising four input neurons and one output neuron. It is established that the final state of this NCN is independent of its initial state and determined entirely by the conditions of learning (sequence of spikes). Dependence of the result of learning on the threshold current of output neuron has been studied. The obtained results open prospects for creating autonomous NCNs capable of being trained to solve complex cognitive tasks.

[1]  Farnood Merrikh-Bayat,et al.  Training and operation of an integrated neuromorphic network based on metal-oxide memristors , 2014, Nature.

[2]  A S Spinelli,et al.  Memristive neural network for on-line learning and tracking with brain-inspired spike timing dependent plasticity , 2017, Scientific Reports.

[3]  Sumio Hosaka,et al.  Emulating the paired-pulse facilitation of a biological synapse with a NiOx-based memristor , 2013 .

[4]  Davood Shahrjerdi,et al.  A sub-1-volt analog metal oxide memristive-based synaptic device with large conductance change for energy-efficient spike-based computing systems , 2016, 1603.03979.

[5]  Narayan Srinivasa,et al.  A functional hybrid memristor crossbar-array/CMOS system for data storage and neuromorphic applications. , 2012, Nano letters.

[6]  Farnood Merrikh-Bayat,et al.  Self-Adaptive Spike-Time-Dependent Plasticity of Metal-Oxide Memristors , 2015, Scientific Reports.

[7]  Thomas Mikolajick,et al.  Rectifying filamentary resistive switching in ion-exfoliated LiNbO3 thin films , 2016 .

[8]  Mirko Hansen,et al.  Unsupervised Hebbian learning experimentally realized with analogue memristive crossbar arrays , 2018, Scientific Reports.

[9]  Peng Lin,et al.  Fully memristive neural networks for pattern classification with unsupervised learning , 2018 .

[10]  Jiaming Zhang,et al.  Analogue signal and image processing with large memristor crossbars , 2017, Nature Electronics.

[11]  Fei Zhuge,et al.  Memristive Synapses for Brain‐Inspired Computing , 2019, Advanced Materials Technologies.

[12]  K. E. Nikiruy,et al.  A Precise Algorithm of Memristor Switching to a State with Preset Resistance , 2018 .

[13]  P. E. Ovchinnikov,et al.  Formation of Weighting Coefficients in an Artificial Neural Network Based on the Memristive Effect in Metal–Oxide–Metal Nanostructures , 2018, Journal of Communications Technology and Electronics.

[14]  Ali Khiat,et al.  Unsupervised learning in probabilistic neural networks with multi-state metal-oxide memristive synapses , 2016, Nature Communications.

[15]  Everton J. Agnes,et al.  Inhibitory Plasticity: Balance, Control, and Codependence. , 2017, Annual review of neuroscience.

[16]  A. V. Emelyanov,et al.  Spike-timing-dependent plasticity of polyaniline-based memristive element , 2018 .

[17]  Victor Erokhin,et al.  First steps towards the realization of a double layer perceptron based on organic memristive devices , 2016 .

[18]  U. Rajendra Acharya,et al.  Deep convolutional neural network for the automated detection and diagnosis of seizure using EEG signals , 2017, Comput. Biol. Medicine.

[19]  V. Rylkov,et al.  Transport, Magnetic, and Memristive Properties of a Nanogranular (CoFeB)x(LiNbOy)100–x Composite Material , 2018 .

[20]  Christophe Loyez,et al.  A 4-fJ/Spike Artificial Neuron in 65 nm CMOS Technology , 2017, Front. Neurosci..

[21]  Daniele Ielmini,et al.  Resistive switching memories based on metal oxides: mechanisms, reliability and scaling , 2016 .