Co-design of DNN model optimization for binary ReRAM array in-memory processing

Yue Guan, Takashi Ohsawa

研究成果: Conference contribution

抄録

In this work, a novel design of ReRAM neuromorphic system is proposed to process deep neural network (DNN) fully in array efficiently. A binary neural network model is constructed and optimized on MNIST dataset. The obtained model is simulated to be processed with the proposed ReRAM array. Co-design between hardware and network model optimization in software is analyzed to achieve feasible hardware design and generalizable model. Deployed with such co-design model, ReRAM array processes DNN with high robustness against fabrication fluctuation.

元の言語English
ホスト出版物のタイトル2019 IEEE 11th International Memory Workshop, IMW 2019
出版者Institute of Electrical and Electronics Engineers Inc.
ISBN(電子版)9781728109817
DOI
出版物ステータスPublished - 2019 5 1
イベント11th IEEE International Memory Workshop, IMW 2019 - Montterey, United States
継続期間: 2019 5 122019 5 15

出版物シリーズ

名前2019 IEEE 11th International Memory Workshop, IMW 2019

Conference

Conference11th IEEE International Memory Workshop, IMW 2019
United States
Montterey
期間19/5/1219/5/15

Fingerprint

Data storage equipment
Processing
Hardware
RRAM
Deep neural networks
Neural networks
Fabrication

ASJC Scopus subject areas

  • Hardware and Architecture
  • Electrical and Electronic Engineering

これを引用

Guan, Y., & Ohsawa, T. (2019). Co-design of DNN model optimization for binary ReRAM array in-memory processing. : 2019 IEEE 11th International Memory Workshop, IMW 2019 [8739722] (2019 IEEE 11th International Memory Workshop, IMW 2019). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/IMW.2019.8739722

Co-design of DNN model optimization for binary ReRAM array in-memory processing. / Guan, Yue; Ohsawa, Takashi.

2019 IEEE 11th International Memory Workshop, IMW 2019. Institute of Electrical and Electronics Engineers Inc., 2019. 8739722 (2019 IEEE 11th International Memory Workshop, IMW 2019).

研究成果: Conference contribution

Guan, Y & Ohsawa, T 2019, Co-design of DNN model optimization for binary ReRAM array in-memory processing. : 2019 IEEE 11th International Memory Workshop, IMW 2019., 8739722, 2019 IEEE 11th International Memory Workshop, IMW 2019, Institute of Electrical and Electronics Engineers Inc., 11th IEEE International Memory Workshop, IMW 2019, Montterey, United States, 19/5/12. https://doi.org/10.1109/IMW.2019.8739722
Guan Y, Ohsawa T. Co-design of DNN model optimization for binary ReRAM array in-memory processing. : 2019 IEEE 11th International Memory Workshop, IMW 2019. Institute of Electrical and Electronics Engineers Inc. 2019. 8739722. (2019 IEEE 11th International Memory Workshop, IMW 2019). https://doi.org/10.1109/IMW.2019.8739722
Guan, Yue ; Ohsawa, Takashi. / Co-design of DNN model optimization for binary ReRAM array in-memory processing. 2019 IEEE 11th International Memory Workshop, IMW 2019. Institute of Electrical and Electronics Engineers Inc., 2019. (2019 IEEE 11th International Memory Workshop, IMW 2019).
@inproceedings{2e637885bd9a40e0b725bd72e95e52d8,
title = "Co-design of DNN model optimization for binary ReRAM array in-memory processing",
abstract = "In this work, a novel design of ReRAM neuromorphic system is proposed to process deep neural network (DNN) fully in array efficiently. A binary neural network model is constructed and optimized on MNIST dataset. The obtained model is simulated to be processed with the proposed ReRAM array. Co-design between hardware and network model optimization in software is analyzed to achieve feasible hardware design and generalizable model. Deployed with such co-design model, ReRAM array processes DNN with high robustness against fabrication fluctuation.",
keywords = "binary neural network, fabrication fluctuation, neuromorphic ReRAM",
author = "Yue Guan and Takashi Ohsawa",
year = "2019",
month = "5",
day = "1",
doi = "10.1109/IMW.2019.8739722",
language = "English",
series = "2019 IEEE 11th International Memory Workshop, IMW 2019",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
booktitle = "2019 IEEE 11th International Memory Workshop, IMW 2019",

}

TY - GEN

T1 - Co-design of DNN model optimization for binary ReRAM array in-memory processing

AU - Guan, Yue

AU - Ohsawa, Takashi

PY - 2019/5/1

Y1 - 2019/5/1

N2 - In this work, a novel design of ReRAM neuromorphic system is proposed to process deep neural network (DNN) fully in array efficiently. A binary neural network model is constructed and optimized on MNIST dataset. The obtained model is simulated to be processed with the proposed ReRAM array. Co-design between hardware and network model optimization in software is analyzed to achieve feasible hardware design and generalizable model. Deployed with such co-design model, ReRAM array processes DNN with high robustness against fabrication fluctuation.

AB - In this work, a novel design of ReRAM neuromorphic system is proposed to process deep neural network (DNN) fully in array efficiently. A binary neural network model is constructed and optimized on MNIST dataset. The obtained model is simulated to be processed with the proposed ReRAM array. Co-design between hardware and network model optimization in software is analyzed to achieve feasible hardware design and generalizable model. Deployed with such co-design model, ReRAM array processes DNN with high robustness against fabrication fluctuation.

KW - binary neural network

KW - fabrication fluctuation

KW - neuromorphic ReRAM

UR - http://www.scopus.com/inward/record.url?scp=85068349191&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85068349191&partnerID=8YFLogxK

U2 - 10.1109/IMW.2019.8739722

DO - 10.1109/IMW.2019.8739722

M3 - Conference contribution

AN - SCOPUS:85068349191

T3 - 2019 IEEE 11th International Memory Workshop, IMW 2019

BT - 2019 IEEE 11th International Memory Workshop, IMW 2019

PB - Institute of Electrical and Electronics Engineers Inc.

ER -