The viability of analog-based accelerators for neuromorphic computing: a survey

Abstract

Focus in deep neural network hardware research for reducing latencies of memory fetches has steered in the direction of analog-based artificial neural networks (ANN). The promise of decreased latencies, increased computational parallelism, and higher storage densities with crossbar non-volatile memory (NVM) based in-memory-computing/processing-in-memory techniques is not without its caveats. This paper surveys this rich landscape and highlights the advantages and challenges of emerging NVMs as multi-level synaptic emulators in various neural network types and applications. Current and potential methods for reliably programming these devices in a crossbar matrix are discussed, as well as techniques for reliably integrating and propagating matrix products to emulate the well-known MAC-like operations throughout the neural network. This paper complements previous surveys, but most importantly uncovers further areas of ongoing research relating to the viability of analog-based ANN implementations based on state-of-the-art NVM technologies in the context of hardware accelerators. While many previous reviews of analog-based ANN focus on device characteristics, this review presents the perspective of crossbar arrays, peripheral circuitry and the required architectural and system considerations for an emerging memory crossbar neural network.