This book describes an extensive and consistent soft error assessment of convolutional neural network (CNN) models from different domains through more than 14.8 million fault injections, considering different precision bit-width configurations, optimization parameters, and processor models. The authors also evaluate the relative performance, memory utilization, and soft error reliability trade-offs analysis of different CNN models considering a compiler-based technique w.r.t. traditional redundancy approaches.
Produkteigenschaften
- Artikelnummer: 9783031185984
- Medium: Buch
- ISBN: 978-3-031-18598-4
- Verlag: Springer Nature Switzerland
- Erscheinungstermin: 02.01.2023
- Sprache(n): Englisch
- Auflage: 1. Auflage 2023
- Serie: Synthesis Lectures on Engineering, Science, and Technology
- Produktform: Gebunden
- Gewicht: 467 g
- Seiten: 131
- Format (B x H x T): 173 x 246 x 14 mm
- Ausgabetyp: Kein, Unbekannt