AdAM: Adaptive Fault-Tolerant Approximate Multiplier for Edge DNN Accelerators
Abstract
In this paper, we propose an architecture of a novel adaptive fault-tolerant approximate multiplier tailored for ASIC-based DNN accelerators.
- Publication:
-
arXiv e-prints
- Pub Date:
- March 2024
- DOI:
- arXiv:
- arXiv:2403.02936
- Bibcode:
- 2024arXiv240302936T
- Keywords:
-
- Computer Science - Artificial Intelligence;
- Computer Science - Hardware Architecture;
- Computer Science - Machine Learning