STT-BSNN: An In-Memory Deep Binary Spiking Neural Network Based on STT-MRAM

Nguyen, Van‐Tinh; Trinh, Quang-Kien; Zhang, Renyuan; Nakashima, Yasuhiko

doi:10.1109/access.2021.3125685

Cited by 15 publications

(19 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We also reported in the table the accuracy when testing with MNIST and FashionMNIST, though, with those simple datasets, the accuracy level is almost saturated and could not be representative. For CIFAR-10, compared to our previous work in Nguyen et al, 12 this work shows similar accuracy with marginal degradation of 1.84%. However, the proposed circuit outperforms in all other aspects, including 6.9% better synaptic energy consumption, ‡ 1.6Â smaller in neuron area, and $1.7Â faster in spiking rate.…”

Section: Effect Of Process Variations and Circuit Nonidealitiessupporting

confidence: 74%

“…This work adopts the SNN model training with binarized weights and rate encoding method. 12 The weights are binarized from full precision values as presented in the BNN model in Rastegari et al 14 In the inference, the IF neuron with batch normalization (BN) can be calculated as follows:…”

Section: Xnor-based Bsnn Inference Modelmentioning

confidence: 99%

“…5 Recently, several implementations of binary spiking neural network (BSNN) based on spintronics memory have been proposed, aiming to overcome drawbacks related to poor computation accuracy due to process variation. [6][7][8][9][10][11][12][13] Spiking neural networks (SNNs) naturally exhibit better fault tolerance capabilities. 6,12 Indeed, in SNN, the inaccuracy due to device mismatch and fluctuation of operating conditions may affect individual spikes (e.g., the spike position) but have little impact on the overall value, which typically is represented by the number of spikes over a fixed period.…”

Section: Introductionmentioning

confidence: 99%

“…[6][7][8][9][10][11][12][13] Spiking neural networks (SNNs) naturally exhibit better fault tolerance capabilities. 6,12 Indeed, in SNN, the inaccuracy due to device mismatch and fluctuation of operating conditions may affect individual spikes (e.g., the spike position) but have little impact on the overall value, which typically is represented by the number of spikes over a fixed period. 12 In other words, SNN pays the tradeoff of longer computation time to gain the desirable computation accuracy.…”

Section: Introductionmentioning

confidence: 99%

“…6,12 Indeed, in SNN, the inaccuracy due to device mismatch and fluctuation of operating conditions may affect individual spikes (e.g., the spike position) but have little impact on the overall value, which typically is represented by the number of spikes over a fixed period. 12 In other words, SNN pays the tradeoff of longer computation time to gain the desirable computation accuracy. All spintronic MTJ-based spiking neuron for BSNN was studied in Wu et al 7,8 In these studies, the synaptic current was converted into probabilistic Poisson events by back-hopping oscillation in an MTJ.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

A low‐power charge‐based integrate‐and‐fire circuit for binarized‐spiking neural network

Duong

Quang

Nguyen

et al. 2023

Circuit Theory & Apps

Self Cite

View full text Add to dashboard Cite

Summary This paper presents a charge‐based integrate‐and‐fire (IF) circuit for in‐memory binary spiking neural networks (BSNNs). The proposed IF circuit can mimic both addition and subtraction operations that permit better incorporation with in‐memory XNOR‐based synapses to implement the BSNN processing core. To evaluate the proposed design, we have developed a framework that incorporates the circuit's imperfections effects into the system‐level simulation. The array circuits use 2T‐2J Spin‐Transfer‐Torque Magnetoresistive RAM (STT‐MRAM) based on a 65‐nm commercial CMOS and a fitted magnetic tunnel junction (MTJ). The system model has been described in Pytorch to best fit the extracted parameters from circuit levels, including the cover of device nonidealities and process variations. The simulation results show that the proposed design can achieve a performance of 5.10 fJ/synapse and reaches 82.01% classification accuracy for CIFAR‐10 under process variation.

show abstract

Section: Effect Of Process Variations and Circuit Nonidealitiessupporting

confidence: 74%

Section: Xnor-based Bsnn Inference Modelmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

A low‐power charge‐based integrate‐and‐fire circuit for binarized‐spiking neural network

Duong

Quang

Nguyen

et al. 2023

Circuit Theory & Apps

Self Cite

View full text Add to dashboard Cite

show abstract

Hardware Implementation for Spiking Neural Networks on Edge Devices

Nguyen

Veeravalli

Fong

2022

Predictive Analytics in Cloud, Fog, and Edge Computing

View full text Add to dashboard Cite

Unconventional computing based on magnetic tunnel junction

Cai

Xin

et al. 2023

Appl. Phys. A

View full text Add to dashboard Cite

The conventional computing method based on the von Neumann architecture is limited by a series of problems such as high energy consumption, finite data exchange bandwidth between processors and storage media, etc., and it is difficult to achieve higher computing efficiency. A more efficient unconventional computing architecture is urgently needed to overcome these problems. Neuromorphic computing and stochastic computing have been considered to be two competitive candidates for unconventional computing, due to their extraordinary potential for energy-efficient and high-performance computing. Although conventional electronic devices can mimic the topology of the human brain, these require high power consumption and large area. Spintronic devices represented by magnetic tunnel junctions (MTJs) exhibit remarkable high-energy efficiency, non-volatility, and similarity to biological nervous systems, making them one of the promising candidates for unconventional computing. In this work, we review the fundamentals of MTJs as well as the development of MTJ-based neurons, synapses, and probabilistic-bit. In the section on neuromorphic computing, we review a variety of neural networks composed of MTJ-based neurons and synapses, including multilayer perceptrons, convolutional neural networks, recurrent neural networks, and spiking neural networks, which are the closest to the biological neural system. In the section on stochastic computing, we review the applications of MTJ-based p-bits, including Boltzmann machines, Ising machines, and Bayesian networks. Furthermore, the challenges to developing these novel technologies are briefly discussed at the end of each section.

show abstract

STT-BSNN: An In-Memory Deep Binary Spiking Neural Network Based on STT-MRAM

Cited by 15 publications

References 32 publications

A low‐power charge‐based integrate‐and‐fire circuit for binarized‐spiking neural network

A low‐power charge‐based integrate‐and‐fire circuit for binarized‐spiking neural network

Hardware Implementation for Spiking Neural Networks on Edge Devices

Unconventional computing based on magnetic tunnel junction

Contact Info

Product

Resources

About