Context-based error correction scheme using recurrent neural network for resilient and efficient intra-chip data transmission

Naoto Sugaya; Masanori Natsui; Takahiro Hanyu

doi:10.1109/ISMVL.2016.42

Context-based error correction scheme using recurrent neural network for resilient and efficient intra-chip data transmission

Naoto Sugaya, Masanori Natsui, Takahiro Hanyu

Systems & Software Division

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

3 Citations (Scopus)

Abstract

An error correction scheme utilizing a brain-inspired learning algorithm, called Recurrent Neural Network (RNN), is proposed for resilient and efficient intra-chip data transmission. RNN has a feature to find partially-clustered time-series data stream from an input data stream and predict the next input data from previous input data stream, which can be utilized for realizing an error correction corresponding to the "context" of the data stream. Through the evaluation of intra-chip data transmission in a general-purpose 32-bit microprocessor, it is demonstrated that the proposed scheme performs 95.9% error reduction with 2-times better data transfer efficiency and 94.2% error reduction with 4-times better data transfer efficiency compared with a conventional error correction scheme.

Original language	English
Title of host publication	Proceedings - 2016 IEEE 46th International Symposium on Multiple-Valued Logic, ISMVL 2016
Publisher	IEEE Computer Society
Pages	72-77
Number of pages	6
ISBN (Electronic)	9781467394888
DOIs	https://doi.org/10.1109/ISMVL.2016.42
Publication status	Published - 2016 Jul 18
Event	46th IEEE International Symposium on Multiple-Valued Logic, ISMVL 2016 - Sapporo, Hokkaido, Japan Duration: 2016 May 18 → 2016 May 20

Publication series

Name	Proceedings of The International Symposium on Multiple-Valued Logic
Volume	2016-July
ISSN (Print)	0195-623X

Other

Other	46th IEEE International Symposium on Multiple-Valued Logic, ISMVL 2016
Country/Territory	Japan
City	Sapporo, Hokkaido
Period	16/5/18 → 16/5/20

Keywords

context-based error correction
deep learning
intelligent information processing
recurrent neural network

ASJC Scopus subject areas

Computer Science(all)
Mathematics(all)

Access to Document

10.1109/ISMVL.2016.42

Cite this

Sugaya, N., Natsui, M., & Hanyu, T. (2016). Context-based error correction scheme using recurrent neural network for resilient and efficient intra-chip data transmission. In Proceedings - 2016 IEEE 46th International Symposium on Multiple-Valued Logic, ISMVL 2016 (pp. 72-77). [7515525] (Proceedings of The International Symposium on Multiple-Valued Logic; Vol. 2016-July). IEEE Computer Society. https://doi.org/10.1109/ISMVL.2016.42

Context-based error correction scheme using recurrent neural network for resilient and efficient intra-chip data transmission. / Sugaya, Naoto; Natsui, Masanori ; Hanyu, Takahiro.
Proceedings - 2016 IEEE 46th International Symposium on Multiple-Valued Logic, ISMVL 2016. IEEE Computer Society, 2016. p. 72-77 7515525 (Proceedings of The International Symposium on Multiple-Valued Logic; Vol. 2016-July).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Sugaya, N, Natsui, M & Hanyu, T 2016, Context-based error correction scheme using recurrent neural network for resilient and efficient intra-chip data transmission. in Proceedings - 2016 IEEE 46th International Symposium on Multiple-Valued Logic, ISMVL 2016., 7515525, Proceedings of The International Symposium on Multiple-Valued Logic, vol. 2016-July, IEEE Computer Society, pp. 72-77, 46th IEEE International Symposium on Multiple-Valued Logic, ISMVL 2016, Sapporo, Hokkaido, Japan, 16/5/18. https://doi.org/10.1109/ISMVL.2016.42

Sugaya N, Natsui M , Hanyu T. Context-based error correction scheme using recurrent neural network for resilient and efficient intra-chip data transmission. In Proceedings - 2016 IEEE 46th International Symposium on Multiple-Valued Logic, ISMVL 2016. IEEE Computer Society. 2016. p. 72-77. 7515525. (Proceedings of The International Symposium on Multiple-Valued Logic). doi: 10.1109/ISMVL.2016.42

Sugaya, Naoto ; Natsui, Masanori ; Hanyu, Takahiro. / Context-based error correction scheme using recurrent neural network for resilient and efficient intra-chip data transmission. Proceedings - 2016 IEEE 46th International Symposium on Multiple-Valued Logic, ISMVL 2016. IEEE Computer Society, 2016. pp. 72-77 (Proceedings of The International Symposium on Multiple-Valued Logic).

@inproceedings{25410b57d2c7433aa7e39e4f37877015,

title = "Context-based error correction scheme using recurrent neural network for resilient and efficient intra-chip data transmission",

abstract = "An error correction scheme utilizing a brain-inspired learning algorithm, called Recurrent Neural Network (RNN), is proposed for resilient and efficient intra-chip data transmission. RNN has a feature to find partially-clustered time-series data stream from an input data stream and predict the next input data from previous input data stream, which can be utilized for realizing an error correction corresponding to the {"}context{"} of the data stream. Through the evaluation of intra-chip data transmission in a general-purpose 32-bit microprocessor, it is demonstrated that the proposed scheme performs 95.9% error reduction with 2-times better data transfer efficiency and 94.2% error reduction with 4-times better data transfer efficiency compared with a conventional error correction scheme.",

keywords = "context-based error correction, deep learning, intelligent information processing, recurrent neural network",

author = "Naoto Sugaya and Masanori Natsui and Takahiro Hanyu",

note = "Publisher Copyright: {\textcopyright} 2016 IEEE.; 46th IEEE International Symposium on Multiple-Valued Logic, ISMVL 2016 ; Conference date: 18-05-2016 Through 20-05-2016",

year = "2016",

month = jul,

day = "18",

doi = "10.1109/ISMVL.2016.42",

language = "English",

series = "Proceedings of The International Symposium on Multiple-Valued Logic",

publisher = "IEEE Computer Society",

pages = "72--77",

booktitle = "Proceedings - 2016 IEEE 46th International Symposium on Multiple-Valued Logic, ISMVL 2016",

}

TY - GEN

T1 - Context-based error correction scheme using recurrent neural network for resilient and efficient intra-chip data transmission

AU - Sugaya, Naoto

AU - Natsui, Masanori

AU - Hanyu, Takahiro

PY - 2016/7/18

Y1 - 2016/7/18

N2 - An error correction scheme utilizing a brain-inspired learning algorithm, called Recurrent Neural Network (RNN), is proposed for resilient and efficient intra-chip data transmission. RNN has a feature to find partially-clustered time-series data stream from an input data stream and predict the next input data from previous input data stream, which can be utilized for realizing an error correction corresponding to the "context" of the data stream. Through the evaluation of intra-chip data transmission in a general-purpose 32-bit microprocessor, it is demonstrated that the proposed scheme performs 95.9% error reduction with 2-times better data transfer efficiency and 94.2% error reduction with 4-times better data transfer efficiency compared with a conventional error correction scheme.

AB - An error correction scheme utilizing a brain-inspired learning algorithm, called Recurrent Neural Network (RNN), is proposed for resilient and efficient intra-chip data transmission. RNN has a feature to find partially-clustered time-series data stream from an input data stream and predict the next input data from previous input data stream, which can be utilized for realizing an error correction corresponding to the "context" of the data stream. Through the evaluation of intra-chip data transmission in a general-purpose 32-bit microprocessor, it is demonstrated that the proposed scheme performs 95.9% error reduction with 2-times better data transfer efficiency and 94.2% error reduction with 4-times better data transfer efficiency compared with a conventional error correction scheme.

KW - context-based error correction

KW - deep learning

KW - intelligent information processing

KW - recurrent neural network

UR - http://www.scopus.com/inward/record.url?scp=84981309680&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84981309680&partnerID=8YFLogxK

U2 - 10.1109/ISMVL.2016.42

DO - 10.1109/ISMVL.2016.42

M3 - Conference contribution

AN - SCOPUS:84981309680

T3 - Proceedings of The International Symposium on Multiple-Valued Logic

SP - 72

EP - 77

BT - Proceedings - 2016 IEEE 46th International Symposium on Multiple-Valued Logic, ISMVL 2016

PB - IEEE Computer Society

T2 - 46th IEEE International Symposium on Multiple-Valued Logic, ISMVL 2016

Y2 - 18 May 2016 through 20 May 2016

ER -

Context-based error correction scheme using recurrent neural network for resilient and efficient intra-chip data transmission

Abstract

Publication series

Other

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this