1/f noise in amorphous Sb2Te3 for energy-efficient stochastic synapses in neuromorphic computing. (22nd October 2021)

• Record Type: Journal Article
• Authors: Kang, Deokyoung; Jang, Suyeon; Choi, Sejeung; Kim, Sangbum
• Is Part Of: Semiconductor science and technology. Volume 36:Number 12(2021)
• Publication Date: 2021-10-22
• Subjects: phase change memory -- Sb2Te3 -- neuromorphic computing -- stochastic synapse -- 1/f noise -- two-level states; Semiconductors -- Periodicals; 621.38152
• DOI: 10.1088/1361-6641/ac251c
• Languages: English
• ISSNs: 0268-1242

Abstract

Abstract: Recent studies on neuromorphic computing have used stochastic synapses to implement power-efficient stochastic computing inspired by unreliable connections between neurons, such as the blank-out noise in the Synaptic Sampling Machine. In this paper, we propose to generate stochasticity by exploiting intrinsic 1 / f noise in phase change memory (PCM) as a synaptic device, negating additional stochastic devices and circuits that deteriorate the synaptic footprints and power consumption. As existing models are limited to demonstrating the spectral density of 1 / f noise, we devised a new model based on two-level state theory with a cutoff frequency, resulting in accurate quantification of the finite normalized variance of current ( σ I 2 / I 2 ) of PCM and its dependence on the volume of the cell and the frequency range in which the measurement is taken. We experimentally verified our model by measuring 1 / f noise of a phase change bridge cell with an as-deposited amorphous Sb2 Te3 as the phase change material. We further analyzed whether the noise in PCM can implement a restricted Boltzmann machine (RBM), in which stochasticity plays a key role, to allow efficient neuromorphic computing. We devised and simulated a spiking neural network (SNN)-based RBM system based on a 832 × 832 PCM synapse array with the intrinsic noise model. As we modeled the normalized synaptic noise with a normal distribution, N μ = 1, σ 2 and optimal standard deviation between 0.01 and 0.05, the on-chip learning and inference test result showed comparable MNIST accuracy and ∼60 times larger estimated energy efficiency than that of the SNN-based RBM, the stochasticity of which is implemented with power-consuming random walk neuron circuits, demonstrating that the intrinsic 1 / f noise of PCM is not a nuisance but an asset to implement efficient neuromorphic computing systems.