Matsukatova Anna N, Iliasov Aleksandr I, Nikiruy Kristina E, Kukueva Elena V, Vasiliev Aleksandr L, Goncharov Boris V, Sitnikov Aleksandr V, Zanaveskin Maxim L, Bugaev Aleksandr S, Demin Vyacheslav A, Rylkov Vladimir V, Emelyanov Andrey V
National Research Center "Kurchatov Institute", 123182 Moscow, Russia.
Faculty of Physics, Lomonosov Moscow State University, 119991 Moscow, Russia.
Nanomaterials (Basel). 2022 Oct 3;12(19):3455. doi: 10.3390/nano12193455.
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)(LiNbO) 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.
卷积神经网络（CNNs）已广泛应用于图像识别和处理任务。基于忆阻器的卷积神经网络积累了新兴忆阻器件的优势，如纳米级关键尺寸、低功耗以及与生物突触的功能相似性。大多数关于基于忆阻器的卷积神经网络的研究要么使用忆阻器的软件模型进行仿真分析，要么实现全硬件卷积神经网络。在此，我们提出一种混合卷积神经网络，它由一个硬件固定的预训练且可解释的特征提取器和一个可训练的软件分类器组成。硬件部分是基于纳米复合材料（Co-Fe-B）（LiNbO）结构的忆阻器无源交叉阵列实现的。构建的双内核卷积神经网络能够以约84%的准确率对二值化的时尚MNIST数据集进行分类。混合卷积神经网络的性能与其他已报道的基于忆阻器的系统相当，而混合卷积神经网络的可训练参数数量则大幅降低。此外，混合卷积神经网络对忆阻特性的变化具有鲁棒性：20%的离散度仅导致准确率下降3%。所获得的结果为基于部分不可靠模拟元件的神经网络的高效可靠实现铺平了道路。
