Challenges in Assessing and Improving Deep Neural Networks’ Reliability
Fernando Fernandes dos Santos, Paolo Rech · IEEE Design and Test · 2024
Editor’s notes: This article addresses reliability assessment challenges for deep neural networks, presents a methodology that combines radiation experiments with software fault simulation, explores hardening solutions to enhance reliability, and concludes with directions for future research. —Jyotika Athavale, Synopsys, USA —Haralampos-G. Stratigopoulos, Sorbonne Université, CNRS, LIP6, France