Unsupervised Online Learning With Multiple Postsynaptic Neurons Based on Spike-Timing-Dependent Plasticity Using a Thin-Film Transistor-Type NOR Flash Memory Array

Abstract

We present a two-layer fully connected neuromorphic system based on a thin-film transistor (TFT)-type NOR flash memory array with multiple postsynaptic (POST) neurons. Unsupervised online learning by spike-timing-dependent plasticity (STDP) on the binary MNIST handwritten datasets is implemented, and its recognition result is determined by measuring firing rate of POST neurons. Using a proposed learning scheme, we investigate the impact of the number of POST neurons in terms of recognition rate. In this neuromorphic system, lateral inhibition function and homeostatic property are exploited for competitive learning of multiple POST neurons. The simulation results demonstrate unsupervised online learning of the full black-and-white MNIST handwritten digits by STDP, which indicates the performance of pattern recognition and classification without preprocessing of input patterns.

Tetrahydroxystilbene Glucoside Ameliorates Infrasound-Induced Central Nervous System (CNS) Injury by Improving Antioxidant and Anti-Inflammatory Capacity

“…Oxidative stress can induce apoptosis [20], and the apoptosis may result in the loss of postsynaptic proteins [21]. Postsynaptic current was modulated by postsynaptic proteins [22] while postsynaptic neurons modulate learning [23] and their dysfunction will impair memory [24]. Thus, oxidative stress can impair learning and memory by affecting postsynaptic activity.…”

Unsupervised online learning of temporal information in spiking neural network using thin-film transistor-type NOR flash memory devices

“…To keep the average threshold voltage of all output neurons, the threshold voltage of neurons which are not fired are lowered by (10/output neuron number) mV. Adaptive neuron threshold is effective in improving the performance of the network by preventing the exclusive or dormant firing of certain neurons [23,24]. A number of previous works were reported to compose CMOS-compatible neuron circuits enabling these various functionalities [25,26], suggesting the possibility of being effectively combined with our synaptic array as shown in figure 5(c).…”

Nanoscale wedge resistive-switching synaptic device and experimental verification of vector-matrix multiplication for hardware neuromorphic application

“…[1][2][3][4][5][6] Researchers incessantly have demonstrated that the more hardware-wisely implemented artificial neural networks (ANNs) are quite successful in tasks related to image pattern recognition. [7][8][9][10] In the development of the ANN hardware, two-terminal electron devices have gained interest due to their geometrical analogy with the biological synapses. [11][12][13][14][15][16][17][18] In this work, resistive-switching random-access memory (RRAM) with high scalability and operation speed is used for a hardware-driven ANN.…”