Improving DNN Hardware Accuracy by In-Memory Computing Noise Injection

Sai Kiran Cherupally; Jian Meng; Adnan Siraj Rakin; Shihui Yin; Mingoo Seok; Deliang Fan; Jae Sun Seo

doi:10.1109/MDAT.2021.3139047

Improving DNN Hardware Accuracy by In-Memory Computing Noise Injection

Sai Kiran Cherupally, Jian Meng, Adnan Siraj Rakin, Shihui Yin, Mingoo Seok, Deliang Fan, Jae Sun Seo

Engineering, Ira A. Fulton Schools of (IAFSE)

Research output: Contribution to journal › Article › peer-review

6 Scopus citations

Abstract

Editor's notes: Like any other designs, in-memory computing (IMC) also suffers from operational inaccuracy induced by hardware noise. In this work, the authors propose to take into account hardware noises during the deep neural network (DNN) training in order to improve the DNN inference accuracy.-Yiran

Original language	English (US)
Pages (from-to)	71-80
Number of pages	10
Journal	IEEE Design and Test
Volume	39
Issue number	4
DOIs	https://doi.org/10.1109/MDAT.2021.3139047
State	Published - Aug 1 2022

Keywords

Deep neural network
Hardware-aware training
In-memory computing
Noise injection

ASJC Scopus subject areas

Software
Hardware and Architecture
Electrical and Electronic Engineering

Access to Document

10.1109/MDAT.2021.3139047

Cite this

@article{57609b60b9f241309bc473d54422ee7d,

title = "Improving DNN Hardware Accuracy by In-Memory Computing Noise Injection",

abstract = "Editor's notes: Like any other designs, in-memory computing (IMC) also suffers from operational inaccuracy induced by hardware noise. In this work, the authors propose to take into account hardware noises during the deep neural network (DNN) training in order to improve the DNN inference accuracy.-Yiran",

keywords = "Deep neural network, Hardware-aware training, In-memory computing, Noise injection",

author = "Cherupally, {Sai Kiran} and Jian Meng and Rakin, {Adnan Siraj} and Shihui Yin and Mingoo Seok and Deliang Fan and Seo, {Jae Sun}",

note = "Publisher Copyright: {\textcopyright} 2013 IEEE.",

year = "2022",

month = aug,

day = "1",

doi = "10.1109/MDAT.2021.3139047",

language = "English (US)",

volume = "39",

pages = "71--80",

journal = "IEEE Design and Test",

issn = "2168-2356",

publisher = "IEEE Computer Society",

number = "4",

}

TY - JOUR

T1 - Improving DNN Hardware Accuracy by In-Memory Computing Noise Injection

AU - Cherupally, Sai Kiran

AU - Meng, Jian

AU - Rakin, Adnan Siraj

AU - Yin, Shihui

AU - Seok, Mingoo

AU - Fan, Deliang

AU - Seo, Jae Sun

PY - 2022/8/1

Y1 - 2022/8/1

N2 - Editor's notes: Like any other designs, in-memory computing (IMC) also suffers from operational inaccuracy induced by hardware noise. In this work, the authors propose to take into account hardware noises during the deep neural network (DNN) training in order to improve the DNN inference accuracy.-Yiran

AB - Editor's notes: Like any other designs, in-memory computing (IMC) also suffers from operational inaccuracy induced by hardware noise. In this work, the authors propose to take into account hardware noises during the deep neural network (DNN) training in order to improve the DNN inference accuracy.-Yiran

KW - Deep neural network

KW - Hardware-aware training

KW - In-memory computing

KW - Noise injection

UR - http://www.scopus.com/inward/record.url?scp=85122310977&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85122310977&partnerID=8YFLogxK

U2 - 10.1109/MDAT.2021.3139047

DO - 10.1109/MDAT.2021.3139047

M3 - Article

AN - SCOPUS:85122310977

SN - 2168-2356

VL - 39

SP - 71

EP - 80

JO - IEEE Design and Test

JF - IEEE Design and Test

IS - 4

ER -

Improving DNN Hardware Accuracy by In-Memory Computing Noise Injection

Abstract

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this