Convolutional Neural Network Based on Crossbar Arrays of (Co-Fe-B)x(LiNbO3)100−x Nanocomposite Memristors

Abstract

Convolutional neural networks (CNNs) have been widely used in image recognition and processing tasks. Memristor-based CNNs accumulate the advantages of emerging memristive devices, such as nanometer critical dimensions, low power consumption, and functional similarity to biological synapses. Most studies on memristor-based CNNs use either software models of memristors for simulation analysis or full hardware CNN realization. Here, we propose a hybrid CNN, consisting of a hardware fixed pre-trained and explainable feature extractor and a trainable software classifier. The hardware part was realized on passive crossbar arrays of memristors based on nanocomposite (Co-Fe-B)x(LiNbO3)100−x structures. The constructed 2-kernel CNN was able to classify the binarized Fashion-MNIST dataset with ~ 84% accuracy. The performance of the hybrid CNN is comparable to the other reported memristor-based systems, while the number of trainable parameters for the hybrid CNN is substantially lower. Moreover, the hybrid CNN is robust to the variations in the memristive characteristics: dispersion of 20% leads to only a 3% accuracy decrease. The obtained results pave the way for the efficient and reliable realization of neural networks based on partially unreliable analog elements.

Adapted MLP-Mixer network based on crossbar arrays of fast and multilevel switching (Co–Fe–B)x(LiNbO3)100−x nanocomposite memristors

“…Memristors with stable multilevel resistive switching can be used in further studies with more complex NCSs capable of learning, such as convolutional networks and others. 25,52,53 During training and inference processes, memristors are switched between different resistive states multiple times, so their immunity to such consecutive switches (endurance) is crucial for NCS operation. Fig.…”

Modulation of polyaniline memristive device switching voltage by nucleotide-free analogue of vitamin B12

“…These devices demonstrate energy-efficient operation in biologically-inspired computing algorithms, such as spiking neural networks [3]. Among other applications, memristive devices are used in formal neuromorphic networks, such as fully connected perceptron [4], reinforcement learning networks [5], convolutional kernels [6], astrocytic networks [7], reservoir computing [8], and macros [9]. Another suitable application for such devices is complementation of biological systems for therapy and improvement of life quality, including neuromorphic vision [10,11], sensing systems [12], and neuroprosthetics [13].…”

An Integrated Lagrangian Modeling Method for Mechanical Systems with Memory Elements

“…Such circuit memory elements and their combination open up new possibilities in electronics. Their applications include (but are not limited to) non-volatile memory, machine learning and neuromorphic computing [16][17][18][19][20][21].…”