CoChat

Soft Error Resilience of Deep Residual Networks for Object Recognition

Open access
Ibrahim, Younis, Wang, Haibin, Bai, Man, Liu, Zhi, Wang, Jianan, Yang, Zhiming, Chen, Zhengming · 2020 · IEEE Access

The DOI matches its Crossref record; 30 of the 31 references it cites were found in a registry or an index, and 1 are not checked yet.

Abstract

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.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1109/access.2020.2968129

Cites

30 of 31 references verified0 citation errors · 0 not checkable

Statistics

Figures

FIGURE 1 ResNet model diagram for ResNet-50, 101, and 152.
FIGURE 1 ResNet model diagram for ResNet-50, 101, and 152.CC-BY · From the PDF
FIGURE 2 Residual module of the ResNet models.
FIGURE 2 Residual module of the ResNet models.CC-BY · From the PDF
FIGURE 3 Basic CUDA-based GPU architecture.
FIGURE 3 Basic CUDA-based GPU architecture.CC-BY · From the PDF
FIGURE 4 Depth influence of the ResNet models for RF modes.
FIGURE 4 Depth influence of the ResNet models for RF modes.CC-BY · From the PDF
FIGURE 5 KVF for each kernel of ResNet model (before hardening). The results obtained by injecting faults into (a) register file, (b) addresses of the instruction outputs, (c) values of the instruction outputs. The whole sub-graph sums up to 100%.
FIGURE 5 KVF for each kernel of ResNet model (before hardening). The results obtained by injecting faults into (a) register file, (b) addresses of the instruction outputs, (c) values of the instruction outputs. The whole sub-graph sums up to 100%.CC-BY · From the PDF
FIGURE 6 Layer vulnerability analysis for each layer of ResNet model (before hardening). The analysis for RF mode injections.
FIGURE 6 Layer vulnerability analysis for each layer of ResNet model (before hardening). The analysis for RF mode injections.CC-BY · From the PDF
FIGURE 8 Layer vulnerability analysis for each layer of ResNet model (before hardening). The analysis for IOV mode injections.
FIGURE 8 Layer vulnerability analysis for each layer of ResNet model (before hardening). The analysis for IOV mode injections.CC-BY · From the PDF
FIGURE 9 KVF for each kernel of ResNet model (after applying our hardening solution). The results obtained by injecting faults into (a) register file, (b) addresses of the instruction outputs, (c) values of the instruction outputs. The whole sub-graph sums up to 100%.
FIGURE 9 KVF for each kernel of ResNet model (after applying our hardening solution). The results obtained by injecting faults into (a) register file, (b) addresses of the instruction outputs, (c) values of the instruction outputs. The whole sub-graph sums up to 100%.CC-BY · From the PDF
Image 9
CC-BY · From the PDF
FIGURE 11 Layer vulnerability evaluation for each layer of ResNet model (after applying hardening solution). The evaluation for IOA mode injections.
FIGURE 11 Layer vulnerability evaluation for each layer of ResNet model (after applying hardening solution). The evaluation for IOA mode injections.CC-BY · From the PDF
FIGURE 12 Layer vulnerability evaluation for each layer of ResNet model (after applying hardening solution). The evaluation for IOV mode injections.
FIGURE 12 Layer vulnerability evaluation for each layer of ResNet model (after applying hardening solution). The evaluation for IOV mode injections.CC-BY · From the PDF
Image 12
CC-BY · From the PDF
Image 13
CC-BY · From the PDF
Image 14
CC-BY · From the PDF
Image 15
CC-BY · From the PDF
Image 16
CC-BY · From the PDF
Image 17
CC-BY · From the PDF
Image 18
CC-BY · From the PDF

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 31 references

Research works printed but not linked: 3

  1. J. Redmon. (2016). Darknet: Open Source Neural Networks in C. [Online]. Available: http://pjreddie.com/darknet/
  2. E. Strohmaier, J. Dongarra, H. Simon, and H. Meuer. (Nov. 2018). The Top 500 List. [Online]. Available: https://www.top500.org/lists/2018/11
  3. NVIDIA Corporation, Santa Clara, CA, USA. NVIDIA GeForce GTX 750 Ti Whitepaper V1.1, [Online]. Available: https://international. download.nvidia.com/geforcecom/international/pdfs/GeForce-GTX-750- TiWhitepaper.pdf

Cited by

See all 82

Metadata

Authors
Ibrahim, Y., Wang, H., Bai, M., Liu, Z., Wang, J., Yang, Z., & Chen, Z.
Affiliations
  • Hohai University (CN)
  • Harbin Institute of Technology (CN)
Venue
IEEE Access
Published
2020
Volume
8
Pages
19490-19503
ISSN
2169-3536 (ISSN-L)
OpenAlex
W3002446690
Cited by
72
References
30
Citations by year
2026: 3 · 2025: 3 · 2024: 23 · 2023: 12 · 2022: 15 · 2021: 13 · 2020: 3
FWCI
3.69
Type
journal article
Language
en
Open access
Locations
  • IEEE Access (publishedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
  • China Institute of Atomic Energy (KFZC2018040205)
  • Changzhou Sci and Tech Program (CZ20180003)
Topics
  • Physical Sciences › Engineering › Electrical and Electronic Engineering › Radiation Effects in Electronics
  • Physical Sciences › Computer Science › Computer Networks and Communications › Distributed systems and fault tolerance
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
Keywords
soft errors, ResNet, object recognition, GPU reliability, radiation-induced faults, fault injection, selective hardening, kernel hardening, safety-critical systems, ResNet50, ResNet-101, ResNet-152, fault tolerance

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Soft Error Resilience of Deep Residual Networks for Object Recognition. https://app.cochat.ai/papers/10.1109/access.2020.2968129
The paper
Ibrahim, Y., Wang, H., Bai, M., Liu, Z., Wang, J., Yang, Z., & Chen, Z. (2020). Soft Error Resilience of Deep Residual Networks for Object Recognition. IEEE Access. https://doi.org/10.1109/access.2020.2968129

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.