Energy-Efficient Convolutional Neural Network Based on Cellular Neural Network Using Beyond-CMOS Technologies

Abstract

In this article, we perform a uniform benchmarking for the convolutional neural network (CoNN) based on the cellular neural network (CeNN) using a variety of beyond-CMOS technologies. Representative charge-based and spintronic device technologies are implemented to enable energy-efficient CeNN related computations. To alleviate the delay and energy overheads of the fully connected layer, a hybrid spintronic CeNN-based CoNN system is proposed. It is shown that low-power FETs and spintronic devices are promising candidates to implement energy-efficient CoNNs based on CeNNs. Specifically, more than 10× improvement in energy-delay product (EDP) is demonstrated for the systems using spin diffusion-based devices and tunneling FETs compared to their conventional CMOS counterparts. INDEX TERMS Beyond-CMOS technology, cellular neural network (CeNN), convolutional neural network (CoNN), spintronics, tunnel FETs (TFETs).

Multiferroic heterostructures for spintronics

“…Spin-based neurons have been proposed and simulated. While showing promise, many rely on magnetic tunnel junctions, which makes readout challenging due to the nature of resistance and spin Hall and spin torques which are unideal due to resistive heating [226].…”