Neurocomputing, volume 407, pages 439-453
Residue Number System-Based Solution for Reducing the Hardware Cost of a Convolutional Neural Network
Chervyakov Nikolay
1
,
Lyakhov Pavel
2, 3
,
Deryabin Maxim
1
,
Nagornov N N
1
,
Valueva Maria
1
,
Valuev Georgii
1
Publication type: Journal Article
Publication date: 2020-09-01
Computer Science Applications
Artificial Intelligence
Cognitive Neuroscience
Abstract
Convolutional neural networks (CNNs) represent deep learning architectures that are currently used in a wide range of applications, including computer vision, speech recognition, time series analysis in finance, and many others. At the same time, CNNs are very demanding in terms of the hardware and time cost of a computing system, which considerably restricts their practical use, e.g., in embedded systems, real-time systems, and mobile volatile devices. The goal of this paper is to reduce the resources required to build and operate CNNs. To achieve this goal, a CNN architecture based on Residue Number System (RNS) and the new Chinese Remainder Theorem with fractions is proposed. The new architecture gives an efficient solution to the main problem of RNSs associated with restoring the number from its residues, which determines the main contribution to the CNN structure. In accordance with the results of hardware simulation on Kintex7 xc7k70tfbg484-2 FPGA, the use of RNS in the convolutional layer of a neural network reduces hardware cost by 32.6% compared to the traditional approach based on the binary number system. In addition, the use of the proposed hardware-software architecture reduces the average image recognition time by 37.06% compared to the software implementation.
Citations by journals
|
1
2
|
|
|
Neurocomputing
|
Neurocomputing
2 publications, 16.67%
|
|
Mathematics and its Applications in New Computer Systems: MANCS-. Lecture Notes in Networks and Systems (424)
|
Mathematics and its Applications in New Computer Systems: MANCS-. Lecture Notes in Networks and Systems (424), 1, 8.33%
Mathematics and its Applications in New Computer Systems: MANCS-. Lecture Notes in Networks and Systems (424)
1 publication, 8.33%
|
|
Applied Sciences (Switzerland)
|
Applied Sciences (Switzerland)
1 publication, 8.33%
|
|
IEEE Access
|
IEEE Access
1 publication, 8.33%
|
|
Computer Optics
|
Computer Optics
1 publication, 8.33%
|
|
Programming and Computer Software
|
Programming and Computer Software
1 publication, 8.33%
|
|
IET Quantum Communication
|
IET Quantum Communication
1 publication, 8.33%
|
|
Journal of Systems and Software
|
Journal of Systems and Software
1 publication, 8.33%
|
|
1
2
|
Citations by publishers
|
1
2
3
|
|
|
Elsevier
|
Elsevier
3 publications, 25%
|
|
IEEE
|
IEEE
2 publications, 16.67%
|
|
Multidisciplinary Digital Publishing Institute (MDPI)
|
Multidisciplinary Digital Publishing Institute (MDPI)
1 publication, 8.33%
|
|
Image Processing Systems Institute of RAS
|
Image Processing Systems Institute of RAS
1 publication, 8.33%
|
|
Pleiades Publishing
|
Pleiades Publishing
1 publication, 8.33%
|
|
Institution of Engineering and Technology (IET)
|
Institution of Engineering and Technology (IET)
1 publication, 8.33%
|
|
1
2
3
|
- We do not take into account publications that without a DOI.
- Statistics recalculated only for publications connected to researchers, organizations and labs registered on the platform.
- Statistics recalculated weekly.
{"yearsCitations":{"type":"bar","data":{"show":true,"labels":[2021,2022,2023,2024],"ids":[0,0,0,0],"codes":[0,0,0,0],"imageUrls":["","","",""],"datasets":[{"label":"Citations number","data":[4,3,4,1],"backgroundColor":["#3B82F6","#3B82F6","#3B82F6","#3B82F6"],"percentage":["33.33","25","33.33","8.33"],"barThickness":null}]},"options":{"indexAxis":"x","maintainAspectRatio":true,"scales":{"y":{"ticks":{"precision":0,"autoSkip":false,"font":{"family":"Montserrat"},"color":"#000000"}},"x":{"ticks":{"stepSize":1,"precision":0,"font":{"family":"Montserrat"},"color":"#000000"}}},"plugins":{"legend":{"position":"top","labels":{"font":{"family":"Montserrat"},"color":"#000000"}},"title":{"display":true,"text":"Citations per year","font":{"size":24,"family":"Montserrat","weight":600},"color":"#000000"}}}},"journals":{"type":"bar","data":{"show":true,"labels":["Neurocomputing","Mathematics and its Applications in New Computer Systems: MANCS-. Lecture Notes in Networks and Systems (424)","Applied Sciences (Switzerland)","IEEE Access","Computer Optics","Programming and Computer Software","IET Quantum Communication","Journal of Systems and Software"],"ids":[4817,35188,16650,25260,1472,7458,35981,9835],"codes":[0,0,0,0,0,0,0,0],"imageUrls":["\/storage\/images\/resized\/GDnYOu1UpMMfMMRV6Aqle4H0YLLsraeD9IP9qScG_medium.webp","","\/storage\/images\/resized\/MjH1ITP7lMYGxeqUZfkt2BnVLgjkk413jwBV97XX_medium.webp","\/storage\/images\/resized\/6scCJegesojp2jubwY3uKCzTAmgsaH2GIFlg6Hfk_medium.webp","\/storage\/images\/resized\/EHGnhmuydZCoMDMFYrH4zhHveWdbfBXM2hIRfmfC_medium.webp","\/storage\/images\/resized\/oZgeErrVFhuDksyqFURLvYS1wtVSBWczh001igGo_medium.webp","\/storage\/images\/resized\/aK6yGjkv01aMMh16lIwFe0Z7wTx4ksfzsYuogiOY_medium.webp","\/storage\/images\/resized\/GDnYOu1UpMMfMMRV6Aqle4H0YLLsraeD9IP9qScG_medium.webp"],"datasets":[{"label":"","data":[2,1,1,1,1,1,1,1],"backgroundColor":["#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6"],"percentage":[16.67,8.33,8.33,8.33,8.33,8.33,8.33,8.33],"barThickness":13}]},"options":{"indexAxis":"y","maintainAspectRatio":false,"scales":{"y":{"ticks":{"precision":0,"autoSkip":false,"font":{"family":"Montserrat"},"color":"#000000"}},"x":{"ticks":{"stepSize":null,"precision":0,"font":{"family":"Montserrat"},"color":"#000000"}}},"plugins":{"legend":{"position":"top","labels":{"font":{"family":"Montserrat"},"color":"#000000"}},"title":{"display":true,"text":"Journals","font":{"size":24,"family":"Montserrat","weight":600},"color":"#000000"}}}},"publishers":{"type":"bar","data":{"show":true,"labels":["Elsevier","IEEE","Multidisciplinary Digital Publishing Institute (MDPI)","Image Processing Systems Institute of RAS","Pleiades Publishing","Institution of Engineering and Technology (IET)"],"ids":[17,6953,202,595,101,117],"codes":[0,0,0,0,0,0],"imageUrls":["\/storage\/images\/resized\/GDnYOu1UpMMfMMRV6Aqle4H0YLLsraeD9IP9qScG_medium.webp","\/storage\/images\/resized\/6scCJegesojp2jubwY3uKCzTAmgsaH2GIFlg6Hfk_medium.webp","\/storage\/images\/resized\/MjH1ITP7lMYGxeqUZfkt2BnVLgjkk413jwBV97XX_medium.webp","\/storage\/images\/resized\/EHGnhmuydZCoMDMFYrH4zhHveWdbfBXM2hIRfmfC_medium.webp","\/storage\/images\/resized\/oZgeErrVFhuDksyqFURLvYS1wtVSBWczh001igGo_medium.webp","\/storage\/images\/resized\/aK6yGjkv01aMMh16lIwFe0Z7wTx4ksfzsYuogiOY_medium.webp"],"datasets":[{"label":"","data":[3,2,1,1,1,1],"backgroundColor":["#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6"],"percentage":[25,16.67,8.33,8.33,8.33,8.33],"barThickness":13}]},"options":{"indexAxis":"y","maintainAspectRatio":false,"scales":{"y":{"ticks":{"precision":0,"autoSkip":false,"font":{"family":"Montserrat"},"color":"#000000"}},"x":{"ticks":{"stepSize":null,"precision":0,"font":{"family":"Montserrat"},"color":"#000000"}}},"plugins":{"legend":{"position":"top","labels":{"font":{"family":"Montserrat"},"color":"#000000"}},"title":{"display":true,"text":"Publishers","font":{"size":24,"family":"Montserrat","weight":600},"color":"#000000"}}}}}
Metrics
Cite this
GOST |
RIS |
BibTex
Cite this
GOST
Copy
Chervyakov N. et al. Residue Number System-Based Solution for Reducing the Hardware Cost of a Convolutional Neural Network // Neurocomputing. 2020. Vol. 407. pp. 439-453.
GOST all authors (up to 50)
Copy
Chervyakov N., Lyakhov P., Deryabin M., Nagornov N. N., Valueva M., Valuev G. Residue Number System-Based Solution for Reducing the Hardware Cost of a Convolutional Neural Network // Neurocomputing. 2020. Vol. 407. pp. 439-453.
Cite this
RIS
Copy
TY - JOUR
DO - 10.1016/j.neucom.2020.04.018
UR - https://doi.org/10.1016%2Fj.neucom.2020.04.018
TI - Residue Number System-Based Solution for Reducing the Hardware Cost of a Convolutional Neural Network
T2 - Neurocomputing
AU - Chervyakov, Nikolay
AU - Lyakhov, Pavel
AU - Deryabin, Maxim
AU - Nagornov, N N
AU - Valueva, Maria
AU - Valuev, Georgii
PY - 2020
DA - 2020/09/01 00:00:00
PB - Elsevier
SP - 439-453
VL - 407
SN - 0925-2312
ER -
Cite this
BibTex
Copy
@article{2020_Chervyakov,
author = {Nikolay Chervyakov and Pavel Lyakhov and Maxim Deryabin and N N Nagornov and Maria Valueva and Georgii Valuev},
title = {Residue Number System-Based Solution for Reducing the Hardware Cost of a Convolutional Neural Network},
journal = {Neurocomputing},
year = {2020},
volume = {407},
publisher = {Elsevier},
month = {sep},
url = {https://doi.org/10.1016%2Fj.neucom.2020.04.018},
pages = {439--453},
doi = {10.1016/j.neucom.2020.04.018}
}