Abstract | ||
---|---|---|
Edge devices must support computationally demanding algorithms, such as neural networks, within tight area/energy budgets. While approximate computing may alleviate these constraints, limiting induced errors remains an open challenge. In this paper, we propose a hardware/software co-design solution via an inexact multiplier, reducing area/power-delay-product requirements by 73/43%, respectively, while still computing exact results when one input is a Fibonacci encoded value. We introduce a retraining strategy to quantize neural network weights to Fibonacci encoded values, ensuring exact computation during inference. We benchmark our strategy on Squeezenet 1.0, DenseNet-121, and ResNet-18, measuring accuracy degradations of only 0.4/1.1/1.7%. |
Year | DOI | Venue |
---|---|---|
2021 | 10.1109/DAC18074.2021.9586245 | 2021 58TH ACM/IEEE DESIGN AUTOMATION CONFERENCE (DAC) |
Keywords | DocType | ISSN |
neural networks, quantization, accelerators, approximate computing | Conference | 0738-100X |
Citations | PageRank | References |
0 | 0.34 | 0 |
Authors | ||
6 |
Name | Order | Citations | PageRank |
---|---|---|---|
william e simon | 1 | 22 | 4.67 |
Valérian Ray | 2 | 0 | 0.34 |
A. Levisse | 3 | 25 | 8.74 |
Giovanni Ansaloni | 4 | 98 | 15.78 |
marina zapater | 5 | 54 | 10.70 |
D. Atienza | 6 | 182 | 24.26 |