Neuromorphic computing using non-volatile memory

By Geoffrey W. Burr, R. Shelby, Abu Sebastian, Sangbum Kim, Seyoung Kim, Severin Sidler, K. Virwani, M. Ishii, Pritish Narayanan, Alessandro Fumarola and 7 more. 2017

1,066 citations recorded by the source corpus.

Abstract Dense crossbar arrays of non-volatile memory (NVM) devices represent one possible path for implementing massively-parallel and highly energy-efficient neuromorphic computing systems. We first review recent advances in the application of NVM devices to three computing paradigms: spiking neural networks (SNNs), deep neural networks (DNNs), and ‘Memcomputing’. In SNNs, NVM synaptic connections are updated by a local learning rule such as spike-timing-dependent-plasticity, a computational approach directly inspired by biology. For DNNs, NVM arrays can represent matrices of synaptic weights, implementing the matrix–vector multiplication needed for algorithms such as backpropagation in an analog yet massively-parallel fashion. This approach could provide significant improvements in power and speed compared to GPU-based DNN training, for applications of commercial significance. We then survey recent research in which different types of NVM devices – including phase change memory, conductive-bridging RAM, filamentary and non-filamentary RRAM, and other NVMs – have been proposed, either as a synapse or as a neuron, for use within a neuromorphic computing application. The relevant virtues and limitations of these devices are assessed, in terms of properties such as conductance dynamic range, (non)linearity and (a)symmetry of conductance response, retention, endurance, required switching power, and device variability. Graphical Abstract

Subfields: Cellular and Molecular Neuroscience, Cognitive Neuroscience, Electrical and Electronic Engineering

View Paper: https://www.semanticscholar.org/paper/41c35fbd3cf42ac544446b2050be4a07a0fc5217

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Cellular and Molecular Neuroscience

In Cognitive Neuroscience

In Electrical and Electronic Engineering

Citations in the corpus

Efficient and self-adaptive in-situ learning in multilayer memristor neural networks (2018, background): “In particular, learning in memristor neural networks has been hampered by signi ﬁ cant statistical variations and ﬂ uctuations in programmed conductance states and the lack of linear and symmetric responses to electric pulses 35 .”
Deep physical neural networks trained with backpropagation (2021, methodology): “Although accurate, large-scale in silico training has been implemented 4–6, 57– Improvements to PAT could extend the utility of PNNs.”
A Survey of Neuromorphic Computing and Neural Networks in Hardware (2017, neutral): “[1384], [1390], [1408], [1439], [2064]–[2069] and neurons”
Memory devices and applications for in-memory computing (2020, background): “NaTuRe NaNoTecHNology This non-volatile accumulative behaviour, in spite of its nonlinear and stochastic nature, can be exploited in several applications, such as training deep neural networks, where the conductance values need to be incrementally modified.”
Organic electronics for neuromorphic computing (2018, background): “…multilayer perceptrons 117 — can be efficiently implemented in hardware, for example in a crossbar array using analogue resistive memory elements 6 , where G n,m is the conductance of the neuromorphic device at the n,m node, V n is applied voltage at the n input, and I m is the read-out current…”
Towards spike-based machine intelligence with neuromorphic computing (2019, background): “Non-volatile technologies 97–103 are usually compared to biological synapses.”
Deep Learning With Spiking Neurons: Opportunities and Challenges (2018, background): “…of STDP is in terms of chip area costly for the presented neuromorphic systems, novel electronic components, so called memristors, may allow for much higher densities of plastic synapses (Jo et al., 2010; Saïghi et al., 2015; Boyn et al., 2017; Burr et al., 2017; Pedretti et al., 2017).”