Soft Error Resilience of Deep Residual Networks for Object Recognition
Convolutional Neural Networks (CNNs) have truly gained attention in object recognition and object classification in particular. When being implemented on Graphics Processing Units (GPUs), deeper networks are more accurate than shallow ones. Residual Networks (ResNets) are one of the deepest CNN arch...
Main Authors: | , , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
IEEE
2020-01-01
|
Series: | IEEE Access |
Subjects: | |
Online Access: | https://ieeexplore.ieee.org/document/8963961/ |
id |
doaj-c1b3af0588d940e292b0ca083f9f23bb |
---|---|
record_format |
Article |
spelling |
doaj-c1b3af0588d940e292b0ca083f9f23bb2021-03-30T02:54:03ZengIEEEIEEE Access2169-35362020-01-018194901950310.1109/ACCESS.2020.29681298963961Soft Error Resilience of Deep Residual Networks for Object RecognitionYounis Ibrahim0Haibin Wang1https://orcid.org/0000-0002-9269-1229Man Bai2Zhi Liu3Jianan Wang4Zhiming Yang5Zhengming Chen6College of IoT Engineering, Hohai University–Changzhou, Changzhou, ChinaCollege of IoT Engineering, Hohai University–Changzhou, Changzhou, ChinaCollege of IoT Engineering, Hohai University–Changzhou, Changzhou, ChinaCollege of IoT Engineering, Hohai University–Changzhou, Changzhou, ChinaNational Key Laboratory of Analog Integrated Circuits, Chongqing, ChinaHarbin Institute of Technology, Harbin, ChinaCollege of IoT Engineering, Hohai University–Changzhou, Changzhou, ChinaConvolutional Neural Networks (CNNs) have truly gained attention in object recognition and object classification in particular. When being implemented on Graphics Processing Units (GPUs), deeper networks are more accurate than shallow ones. Residual Networks (ResNets) are one of the deepest CNN architectures used in various fields including safety-critical ones. GPUs have proven to be the major accelerator for CNN models. However, modern GPUs are prone to radiation-induced soft errors, which is a serious issue in safety-compliant systems. In this work, we analyze and propose an approach to address the reliability of ResNet on GPUs. We firstly analyze three popular ResNet models, explicitly, ResNet-50, ResNet-101, and ResNet-152 through NVIDIA's fault injector, SASSIFI. We perform an in-depth analysis of the model from the perspective of layer and kernel vulnerability. Then, we experimentally show the vulnerability of ResNet models and identify the most vulnerable portions. Finally, we validate our solution, which is a selective-hardening technique, through hardening the worth-hardening kernels to avoid unnecessary overheads. Our strategy is demonstrated to mask up to 93.38% of the injected errors with performance overhead less than 5.35%. Furthermore, the percentage of the errors causing misclassifications can be reduced from 4.2% to 0.104%, thereby significantly improving the model's reliability.https://ieeexplore.ieee.org/document/8963961/Convolutional neural networksresidual networkssafety-critical systemsGPUsreliabilitysoft error |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Younis Ibrahim Haibin Wang Man Bai Zhi Liu Jianan Wang Zhiming Yang Zhengming Chen |
spellingShingle |
Younis Ibrahim Haibin Wang Man Bai Zhi Liu Jianan Wang Zhiming Yang Zhengming Chen Soft Error Resilience of Deep Residual Networks for Object Recognition IEEE Access Convolutional neural networks residual networks safety-critical systems GPUs reliability soft error |
author_facet |
Younis Ibrahim Haibin Wang Man Bai Zhi Liu Jianan Wang Zhiming Yang Zhengming Chen |
author_sort |
Younis Ibrahim |
title |
Soft Error Resilience of Deep Residual Networks for Object Recognition |
title_short |
Soft Error Resilience of Deep Residual Networks for Object Recognition |
title_full |
Soft Error Resilience of Deep Residual Networks for Object Recognition |
title_fullStr |
Soft Error Resilience of Deep Residual Networks for Object Recognition |
title_full_unstemmed |
Soft Error Resilience of Deep Residual Networks for Object Recognition |
title_sort |
soft error resilience of deep residual networks for object recognition |
publisher |
IEEE |
series |
IEEE Access |
issn |
2169-3536 |
publishDate |
2020-01-01 |
description |
Convolutional Neural Networks (CNNs) have truly gained attention in object recognition and object classification in particular. When being implemented on Graphics Processing Units (GPUs), deeper networks are more accurate than shallow ones. Residual Networks (ResNets) are one of the deepest CNN architectures used in various fields including safety-critical ones. GPUs have proven to be the major accelerator for CNN models. However, modern GPUs are prone to radiation-induced soft errors, which is a serious issue in safety-compliant systems. In this work, we analyze and propose an approach to address the reliability of ResNet on GPUs. We firstly analyze three popular ResNet models, explicitly, ResNet-50, ResNet-101, and ResNet-152 through NVIDIA's fault injector, SASSIFI. We perform an in-depth analysis of the model from the perspective of layer and kernel vulnerability. Then, we experimentally show the vulnerability of ResNet models and identify the most vulnerable portions. Finally, we validate our solution, which is a selective-hardening technique, through hardening the worth-hardening kernels to avoid unnecessary overheads. Our strategy is demonstrated to mask up to 93.38% of the injected errors with performance overhead less than 5.35%. Furthermore, the percentage of the errors causing misclassifications can be reduced from 4.2% to 0.104%, thereby significantly improving the model's reliability. |
topic |
Convolutional neural networks residual networks safety-critical systems GPUs reliability soft error |
url |
https://ieeexplore.ieee.org/document/8963961/ |
work_keys_str_mv |
AT younisibrahim softerrorresilienceofdeepresidualnetworksforobjectrecognition AT haibinwang softerrorresilienceofdeepresidualnetworksforobjectrecognition AT manbai softerrorresilienceofdeepresidualnetworksforobjectrecognition AT zhiliu softerrorresilienceofdeepresidualnetworksforobjectrecognition AT jiananwang softerrorresilienceofdeepresidualnetworksforobjectrecognition AT zhimingyang softerrorresilienceofdeepresidualnetworksforobjectrecognition AT zhengmingchen softerrorresilienceofdeepresidualnetworksforobjectrecognition |
_version_ |
1724184452411686912 |