2020
DOI: 10.1109/tcad.2019.2897631
|View full text |Cite
|
Sign up to set email alerts
|

SPINBIS: Spintronics-Based Bayesian Inference System With Stochastic Computing

Abstract: Bayesian inference is an effective approach for solving statistical learning problems, especially with uncertainty and incompleteness. However, Bayesian inference is a computingintensive task whose efficiency is physically limited by the bottlenecks of conventional computing platforms. In this work, a spintronics based stochastic computing approach is proposed for efficient Bayesian inference. The inherent stochastic switching behaviors of spintronic devices are exploited to build stochastic bitstream generato… Show more

Help me understand this report
View preprint versions

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
2
1

Citation Types

1
4
0

Year Published

2020
2020
2024
2024

Publication Types

Select...
7
1
1

Relationship

1
8

Authors

Journals

citations
Cited by 19 publications
(5 citation statements)
references
References 46 publications
1
4
0
Order By: Relevance
“…The bitstream passed all the 15 tests. The Stochastic Computing Correlation (SCC) measurement adopted in SPINBIS [22] yields SCC values within ±0.01 for our bitstreams, which is very close to the ideal value of 0.…”
Section: Randomness Of the Mtj's Behaviorsupporting
confidence: 66%
See 1 more Smart Citation
“…The bitstream passed all the 15 tests. The Stochastic Computing Correlation (SCC) measurement adopted in SPINBIS [22] yields SCC values within ±0.01 for our bitstreams, which is very close to the ideal value of 0.…”
Section: Randomness Of the Mtj's Behaviorsupporting
confidence: 66%
“…Additionally, the MTJ-based random number generator designed in Reference [58] passed 12 out of the 15 NIST SP-800 statistical tests. And the work in Reference [22] tests the suitability of using MTJs for generating Stochastic Numbers and finds low levels of self-and crosscorrelation among the generated bitstreams.…”
Section: Randomness Of the Mtj's Behaviormentioning
confidence: 99%
“…It is implemented by hybrid CMOS/MTJ technologies named spintronic-based Bayesian inference system (SPINBIS) ( Figure 7C ). The aim of proposing the SPINBIS is to enhance the stability of SBG and to use a smaller number of SBGs ( Jia et al, 2020 ). The outcome probability of each SBG is predetermined and is then multiplexed through the switch matrix, which is a crossbar array.…”
Section: New Computing Architecture With Nonvolatile Memory Elements For Bayesian Network Implementationmentioning
confidence: 99%
“…(i) Leakage wall 5,6 , (ii) Reliability wall 4 , and (iii) Cost wall 4,5 . Therefore, to keep the pace with the exploding market needs, novel alternative technologies are under investigation, e.g., graphene [7][8][9][10] , memristors [11][12][13][14][15] , and spintronics [16][17][18][19][20] . Different spintronics technologies have been developed based on, e.g., magnetization switching 21 , Skyrmions generation 22,23 , rectified tunnel magnetoresistance 24 , anomalous Hall effect, and negative differential resistance magnetic tunnel junctions 25 .…”
Section: Introductionmentioning
confidence: 99%