EUVIMEDEUVIMED entdecken
Crossref

ArchRepair : Block-Level Architecture-Oriented Repairing for Deep Neural Networks

Hua Qi, Zhijie Wang, Qing Guo, Jianlang Chen, Felix Juefei-Xu, Fuyuan Zhang, Lei Ma, Jianjun Zhao

ACM Transactions on Software Engineering and Methodology · 30.09.2023 · 32 (5) · 1-31

DOI: 10.1145/3585005 ↗

Kurz gefasst

Fachartikel in der Zeitschrift ACM Transactions on Software Engineering and Methodology, online erschienen am 24.07.2023. Verfasst von Hua Qi und 7 weiteren Autor:innen. Die Arbeit wurde bisher 17-mal zitiert (Crossref) und nennt 76 Literaturangaben. Gefördert von: JST-Mirai Program, JSPS KAKENHI, University Fellowships Toward the Creation of Science Technology Innovation.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Over the past few years, deep neural networks (DNNs) have achieved tremendous success and have been continuously applied in many application domains. However, during the practical deployment in industrial tasks, DNNs are found to be erroneous-prone due to various reasons such as overfitting and lacking of robustness to real-world corruptions during practical usage. To address these challenges, many recent attempts have been made to repair DNNs for version updates under practical operational contexts by updating weights (i.e., network parameters) through retraining, fine-tuning, or direct weight fixing at a neural level. Nevertheless, existing solutions often neglect the effects of neural network architecture and weight relationships across neurons and layers. In this work, as the first attempt, we initiate to repair DNNs by jointly optimizing the architecture and weights at a higher (i.e., block level). We first perform empirical studies to investigate the limitation of whole network-level and layer-level repairing, which motivates us to explore a novel repairing direction for DNN repair at the block level. To this end, we need to further consider techniques to address two key technical challenges, i.e., block localization , where we should localize the targeted block that we need to fix; and how to perform joint architecture and weight repairing . Specifically, we first propose adversarial-aware spectrum analysis for vulnerable block localization that considers the neurons’ status and weights’ gradients in blocks during the forward and backward processes, which enables more accurate candidate block localization for repairing even under a few examples. Then, we further propose the architecture-oriented search-based repairing that relaxes the targeted block to a continuous repairing search space at higher deep feature levels. By jointly optimizing the architecture and weights in that space, we can identify a much better block architecture. We implement our proposed repairing techniques as a tool, named ArchRepair , and conduct extensive experiments to validate the proposed method. The results show that our method can not only repair but also enhance accuracy and robustness, outperforming the state-of-the-art DNN repair techniques.

Förderung

  • JST-Mirai Program
  • JSPS KAKENHI
  • University Fellowships Toward the Creation of Science Technology Innovation
  • Canada CIFAR AI Chairs Program and the Natural Sciences and Engineering Research Council of Canada

Zitieren

APA

Qi, H., Wang, Z., Guo, Q., Chen, J., Juefei-Xu, F., Zhang, F., Ma, L., & Zhao, J. (2023). ArchRepair : Block-Level Architecture-Oriented Repairing for Deep Neural Networks. ACM Transactions on Software Engineering and Methodology, 32(5), 1-31. https://doi.org/10.1145/3585005

Vancouver

Qi H, Wang Z, Guo Q, Chen J, Juefei-Xu F, Zhang F, et al. ArchRepair : Block-Level Architecture-Oriented Repairing for Deep Neural Networks. ACM Transactions on Software Engineering and Methodology. 2023;32(5):1-31. doi:10.1145/3585005

BibTeX
@article{Qi2023,
  author = {Qi, Hua and Wang, Zhijie and Guo, Qing and Chen, Jianlang and Juefei-Xu, Felix and Zhang, Fuyuan and Ma, Lei and Zhao, Jianjun},
  title = {ArchRepair : Block-Level Architecture-Oriented Repairing for Deep Neural Networks},
  journal = {ACM Transactions on Software Engineering and Methodology},
  year = {2023},
  volume = {32},
  number = {5},
  pages = {1--31},
  doi = {10.1145/3585005}
}
Kostenlos starten →