TaOx Synapse Array Based on Ion Profile Engineering for High Accuracy Neuromorpic Computing

Jingjing Yang,Jiadi Zhu,Bingjie Dang,Teng Zhang,Qingxi Duan,Liying Xu,Keqin Liu,Zhiting Lin,Ru Huang,Yuchao Yang

DOI: https://doi.org/10.1109/cstic49141.2020.9282508

2020-01-01

Abstract:Resistive random-access memory is a promising candidate for high-density, low-power neuromorphic computing. Here, we demonstrate a thermally oxidized TaO x RRAM array with built-in oxygen concentration gradient, which ensures good linearity and symmetry and consequently better inference accuracy on both MNIST and CIFAR10 datasets after training with perceptron, LeNet-5 and ResNet-18 networks, thus illustrating the potentiation of applying RRAMs for high-density, high-accuracy, large-scale neuromorphic computing.