EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Adversarial Robustness Enhancement for Deep Learning-Based Soft Sensors: An Adversarial Training Strategy Using Historical Gradients and Domain Adaptation

Runyuan Guo iD, Qingyuan Chen, Han Liu iD, Wenqing Wang iD

Sensors · 17.06.2024 · 24 (12) · 3909

DOI: 10.3390/s24123909 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 17.06.2024. Verfasst von Runyuan Guo und 3 weiteren Autor:innen. Beteiligte Einrichtung: School of Automation and Information Engineering, Xi'an University of Technology, Xi'an 710048, China. Die Arbeit wurde bisher 38-mal zitiert (Crossref) und nennt 32 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Natural Science Foundation of China, Natural Science Basic Research Program of Shaanxi, Doctoral Scientific Research Startup Foundation of Xi'an University of Technology.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Despite their high prediction accuracy, deep learning-based soft sensor (DLSS) models face challenges related to adversarial robustness against malicious adversarial attacks, which hinder their widespread deployment and safe application. Although adversarial training is the primary method for enhancing adversarial robustness, existing adversarial-training-based defense methods often struggle with accurately estimating transfer gradients and avoiding adversarial robust overfitting. To address these issues, we propose a novel adversarial training approach, namely domain-adaptive adversarial training (DAAT). DAAT comprises two stages: historical gradient-based adversarial attack (HGAA) and domain-adaptive training. In the first stage, HGAA incorporates historical gradient information into the iterative process of generating adversarial samples. It considers gradient similarity between iterative steps to stabilize the updating direction, resulting in improved transfer gradient estimation and stronger adversarial samples. In the second stage, a soft sensor domain-adaptive training model is developed to learn common features from adversarial and original samples through domain-adaptive training, thereby avoiding excessive leaning toward either side and enhancing the adversarial robustness of DLSS without robust overfitting. To demonstrate the effectiveness of DAAT, a DLSS model for crystal quality variables in silicon single-crystal growth manufacturing processes is used as a case study. Through DAAT, the DLSS achieves a balance between defense against adversarial samples and prediction accuracy on normal samples to some extent, offering an effective approach for enhancing the adversarial robustness of DLSS.

Schlagwörter

Stichwörter der Autor:innen

adversarial attack and defensedeep learningsoft sensors

Autor:innen und Institutionen

  • School of Automation and Information Engineering, Xi'an University of Technology, Xi'an 710048, China.

ORCID-Profile: Runyuan Guo, Han Liu, Wenqing Wang

Förderung

  • National Natural Science Foundation of China
  • Natural Science Basic Research Program of Shaanxi
  • Doctoral Scientific Research Startup Foundation of Xi'an University of Technology
  • Doctoral Scientific Research Startup Foundation of Xi’an University of Technology

Zitieren

APA

Guo, R., Chen, Q., Liu, H., & Wang, W. (2024). Adversarial Robustness Enhancement for Deep Learning-Based Soft Sensors: An Adversarial Training Strategy Using Historical Gradients and Domain Adaptation. Sensors, 24(12), 3909. https://doi.org/10.3390/s24123909

Vancouver

Guo R, Chen Q, Liu H, Wang W. Adversarial Robustness Enhancement for Deep Learning-Based Soft Sensors: An Adversarial Training Strategy Using Historical Gradients and Domain Adaptation. Sensors. 2024;24(12):3909. doi:10.3390/s24123909

BibTeX
@article{Guo2024,
  author = {Guo, Runyuan and Chen, Qingyuan and Liu, Han and Wang, Wenqing},
  title = {Adversarial Robustness Enhancement for Deep Learning-Based Soft Sensors: An Adversarial Training Strategy Using Historical Gradients and Domain Adaptation},
  journal = {Sensors},
  year = {2024},
  volume = {24},
  number = {12},
  pages = {3909},
  doi = {10.3390/s24123909}
}
Kostenlos starten →