Improving Noise Tolerance of Mixed-Signal Neural Networks

04/02/2019 ∙ by Michael Klachko, et al. ∙ 0

Mixed-signal hardware accelerators for deep learning achieve orders of magnitude better power efficiency than their digital counterparts. In the ultra-low power consumption regime, limited signal precision inherent to analog computation becomes a challenge. We perform a case study of a 6-layer convolutional neural network running on a mixed-signal accelerator and evaluate its sensitivity to hardware specific noise. We apply various methods to improve noise robustness of the network and demonstrate an effective way to optimize useful signal ranges through adaptive signal clipping. The resulting model is robust enough to achieve 80.2 just 1.4 mW power budget, while 6 mW budget allows us to achieve 87.1 accuracy, which is within 1 unoptimized version of the same model achieves only 67.7 and 78.6

READ FULL TEXT
POST COMMENT

Comments

There are no comments yet.

Authors

page 1

page 2

page 3

page 4

Code Repositories

NoisyNet

Code for "Improving Noise Tolerance of Mixed-Signal Neural Networks" https://arxiv.org/abs/1904.01705


view repo
This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.