Robustness Evaluation of Restricted Boltzmann Machine against Memory and Logic Error

Yasushi FUKUDA; Zule XU; Takayuki KAWAHARA

doi:10.1587/transele.E100.C.1118

Robustness Evaluation of Restricted Boltzmann Machine against Memory and Logic Error

Yasushi FUKUDA, Zule XU, Takayuki KAWAHARA

Full Text Views

0

Share
Cite this

Summary :

In an IoT system, neural networks have the potential to perform advanced information processing in various environments. To clarify this, the robustness of a restricted Boltzmann machine (RBM) used for deep neural networks, such as a deep belief network (DBN), was studied in this paper. Even if memory or logic errors occurred in the circuit operating in the RBM while pre-training the DBN, they did not affect the identification rate of the DBN, showing the robustness of the RBM. In addition, robustness against soft errors was evaluated. The soft errors had almost no influence on the RBM unless they were as large as 10¹² times or more in the 50-nm CMOS process.

Publication: IEICE TRANSACTIONS on Electronics Vol.E100-C No.12 pp.1118-1121

Publication Date: 2017/12/01

Publicized

Online ISSN: 1745-1353

DOI: 10.1587/transele.E100.C.1118

Type of Manuscript: BRIEF PAPER

Category: Integrated Electronics

Authors

Yasushi FUKUDA
  Tokyo University of Science
Zule XU
  Tokyo University of Science
Takayuki KAWAHARA
  Tokyo University of Science

Keyword

IoT, neural network, RBM, DBN, soft error

Cite this

Copy

Yasushi FUKUDA, Zule XU, Takayuki KAWAHARA, "Robustness Evaluation of Restricted Boltzmann Machine against Memory and Logic Error" in IEICE TRANSACTIONS on Electronics, vol. E100-C, no. 12, pp. 1118-1121, December 2017, doi: 10.1587/transele.E100.C.1118.
Abstract: In an IoT system, neural networks have the potential to perform advanced information processing in various environments. To clarify this, the robustness of a restricted Boltzmann machine (RBM) used for deep neural networks, such as a deep belief network (DBN), was studied in this paper. Even if memory or logic errors occurred in the circuit operating in the RBM while pre-training the DBN, they did not affect the identification rate of the DBN, showing the robustness of the RBM. In addition, robustness against soft errors was evaluated. The soft errors had almost no influence on the RBM unless they were as large as 10¹² times or more in the 50-nm CMOS process.
URL: https://globals.ieice.org/en_transactions/electronics/10.1587/transele.E100.C.1118/_p

Copy

@ARTICLE{e100-c_12_1118,
author={Yasushi FUKUDA, Zule XU, Takayuki KAWAHARA, },
journal={IEICE TRANSACTIONS on Electronics},
title={Robustness Evaluation of Restricted Boltzmann Machine against Memory and Logic Error},
year={2017},
volume={E100-C},
number={12},
pages={1118-1121},
abstract={In an IoT system, neural networks have the potential to perform advanced information processing in various environments. To clarify this, the robustness of a restricted Boltzmann machine (RBM) used for deep neural networks, such as a deep belief network (DBN), was studied in this paper. Even if memory or logic errors occurred in the circuit operating in the RBM while pre-training the DBN, they did not affect the identification rate of the DBN, showing the robustness of the RBM. In addition, robustness against soft errors was evaluated. The soft errors had almost no influence on the RBM unless they were as large as 10¹² times or more in the 50-nm CMOS process.},
keywords={},
doi={10.1587/transele.E100.C.1118},
ISSN={1745-1353},
month={December},}

Copy

TY - JOUR
TI - Robustness Evaluation of Restricted Boltzmann Machine against Memory and Logic Error
T2 - IEICE TRANSACTIONS on Electronics
SP - 1118
EP - 1121
AU - Yasushi FUKUDA
AU - Zule XU
AU - Takayuki KAWAHARA
PY - 2017
DO - 10.1587/transele.E100.C.1118
JO - IEICE TRANSACTIONS on Electronics
SN - 1745-1353
VL - E100-C
IS - 12
JA - IEICE TRANSACTIONS on Electronics
Y1 - December 2017
AB - In an IoT system, neural networks have the potential to perform advanced information processing in various environments. To clarify this, the robustness of a restricted Boltzmann machine (RBM) used for deep neural networks, such as a deep belief network (DBN), was studied in this paper. Even if memory or logic errors occurred in the circuit operating in the RBM while pre-training the DBN, they did not affect the identification rate of the DBN, showing the robustness of the RBM. In addition, robustness against soft errors was evaluated. The soft errors had almost no influence on the RBM unless they were as large as 10¹² times or more in the 50-nm CMOS process.
ER -