Statistical Training for Neuromorphic Computing using Memristor-based Crossbars Considering Process Variations and Noise
Statistical Training for Neuromorphic Computing using Memristor-based Crossbars Considering Process Variations and Noise
复制标题
DOI：
10.23919/date48585.2020.9116244
复制
发表时间：
2020-03
期刊：
影响因子：
--
通讯作者：
Ying Zhu;Grace Li Zhang;Tianchen Wang;Bing Li;Yiyu Shi;Tsung-Yi Ho;Ulf Schlichtmann
中科院分区：
文献类型：
--
作者：
Ying Zhu;Grace Li Zhang;Tianchen Wang;Bing Li;Yiyu Shi;Tsung-Yi Ho;Ulf Schlichtmann
基于忆阻器的交叉栅是一种很有吸引力的加速神经形态计算的平台。然而，如果不加以解决，制造过程中的工艺变化和忆阻器中的噪声会导致显著的精度损失。在本文中，我们建议将过程变化和噪声作为相关随机变量建模，并在训练过程中将它们纳入成本函数中。因此，经过统计训练后的权重变得更加鲁棒，并结合全局变差补偿提供了稳定的推理精度。仿真结果表明，在两层全连接神经网络中，推理精度的均值和标准差可以显著提高，分别提高54%和31%。
Memristor-based crossbars are an attractive platform to accelerate neuromorphic computing. However, process variations during manufacturing and noise in memristors cause significant accuracy loss if not addressed. In this paper, we propose to model process variations and noise as correlated random variables and incorporate them into the cost function during training. Consequently, the weights after this statistical training become more robust and together with global variation compensation provide a stable inference accuracy. Simulation results demonstrate that the mean value and the standard deviation of the inference accuracy can be improved significantly, by even up to 54% and 31%, respectively, in a two-layer fully connected neural network.