Understanding Error Propagation in Deep-Learning Neural Networks’ Accelerators and Applications
Editor’s notes: This article reviews some of the first experiments on deep neural networks to analyze the propagation of soft errors from hardware to the application level, along with some of the first error mitigation strategies. —Jyotika Athavale, Synopsys, USA —Haralampos-G. Stratigopoulos, Sorbonne Université, CNRS, LIP6, France
Authors
- Joel Emer
- Guanpeng Li (ORCID: https://orcid.org/0000-0001-7773-7826)
- Timothy Tsai (ORCID: https://orcid.org/0000-0001-8190-675X)
- Siva Kumar Sastry Hari (ORCID: https://orcid.org/0000-0001-8346-7981)
- Stephen W. Keckler (ORCID: https://orcid.org/0000-0001-6701-6099)
- Karthik Pattabiraman (ORCID: https://orcid.org/0000-0003-2380-3415)
- Michael B. Sullivan (ORCID: https://orcid.org/0000-0001-6537-2065)
Institutions
- University of Iowa (US)
- University of British Columbia (CA)
- Film Independent (US)
- Nvidia (United States) (US)
Publication Details
- Journal
- IEEE Design and Test
- Published
- 2025-02-21
- DOI
- https://doi.org/10.1109/mdat.2025.3544537
- Citations
- 3
- Primary Topic
- Neural Networks and Applications
- Type
- article
- Field-Weighted Citation Impact
- 7.44