中国物理B ›› 2025, Vol. 34 ›› Issue (4): 46104-046104.doi: 10.1088/1674-1056/adacd1

• • 上一篇    下一篇

Enhancing neural network robustness: Laser fault injection resistance in 55-nm SRAM for space applications

Qing Liu(刘清)1,2, Haomiao Cheng(程浩淼)1,2, Xiang Yao(姚骧)1,2, Zhengxuan Zhang(张正选)1,2, Zhiyuan Hu(胡志远)1,2, and Dawei Bi(毕大炜)1,2,†   

  1. 1 State Key Laboratory of Materials for Integrated Circuits, Shanghai Institute of Microsystem and Information Technology, Shanghai 200050, China;
    2 University of the Chinese Academy of Sciences, Beijing 100049, China
  • 收稿日期:2024-12-19 修回日期:2025-01-20 接受日期:2025-01-22 出版日期:2025-04-15 发布日期:2025-04-15
  • 通讯作者: Dawei Bi E-mail:davidb@mail.sim.ac.cn

Enhancing neural network robustness: Laser fault injection resistance in 55-nm SRAM for space applications

Qing Liu(刘清)1,2, Haomiao Cheng(程浩淼)1,2, Xiang Yao(姚骧)1,2, Zhengxuan Zhang(张正选)1,2, Zhiyuan Hu(胡志远)1,2, and Dawei Bi(毕大炜)1,2,†   

  1. 1 State Key Laboratory of Materials for Integrated Circuits, Shanghai Institute of Microsystem and Information Technology, Shanghai 200050, China;
    2 University of the Chinese Academy of Sciences, Beijing 100049, China
  • Received:2024-12-19 Revised:2025-01-20 Accepted:2025-01-22 Online:2025-04-15 Published:2025-04-15
  • Contact: Dawei Bi E-mail:davidb@mail.sim.ac.cn

PDF (PC)

249

赞

摘要： The integration of artificial intelligence (AI) with satellite technology is ushering in a new era of space exploration, with small satellites playing a pivotal role in advancing this field. However, the deployment of machine learning (ML) models in space faces distinct challenges, such as single event upsets (SEUs), which are triggered by space radiation and can corrupt the outputs of neural networks. To defend against this threat, we investigate laser-based fault injection techniques on 55-nm SRAM cells, aiming to explore the impact of SEUs on neural network performance. In this paper, we propose a novel solution in the form of Bin-DNCNN, a binary neural network (BNN)-based model that significantly enhances robustness to radiation-induced faults. We conduct experiments to evaluate the denoising effectiveness of different neural network architectures, comparing their resilience to weight errors before and after fault injections. Our experimental results demonstrate that binary neural networks (BNNs) exhibit superior robustness to weight errors compared to traditional deep neural networks (DNNs), making them a promising candidate for spaceborne AI applications.

关键词: single event effects, convolutional neural network, fault injection, SRAM

Abstract: The integration of artificial intelligence (AI) with satellite technology is ushering in a new era of space exploration, with small satellites playing a pivotal role in advancing this field. However, the deployment of machine learning (ML) models in space faces distinct challenges, such as single event upsets (SEUs), which are triggered by space radiation and can corrupt the outputs of neural networks. To defend against this threat, we investigate laser-based fault injection techniques on 55-nm SRAM cells, aiming to explore the impact of SEUs on neural network performance. In this paper, we propose a novel solution in the form of Bin-DNCNN, a binary neural network (BNN)-based model that significantly enhances robustness to radiation-induced faults. We conduct experiments to evaluate the denoising effectiveness of different neural network architectures, comparing their resilience to weight errors before and after fault injections. Our experimental results demonstrate that binary neural networks (BNNs) exhibit superior robustness to weight errors compared to traditional deep neural networks (DNNs), making them a promising candidate for spaceborne AI applications.

Key words: single event effects, convolutional neural network, fault injection, SRAM

中图分类号:  (Semiconductors)

  • 61.82.Fk
61.80.Hg (Neutron radiation effects) 07.05.Mh (Neural networks, fuzzy logic, artificial intelligence) 85.35.-p (Nanoelectronic devices)