Separable Binary Convolutional Neural Network on Embedded Systems

doi:10.1109/TC.2020.2973974

Separable Binary Convolutional Neural Network on Embedded Systems

2020-10-01

1

PRE

AI

DOI：10.1109/TC.2020.2973974

Original

Original request for help

Save

Abstract

En 中文

We have witnessed the tremendous success of deep neural networks. However, this success comes with the considerable memory and computational costs which make it difficult to deploy these networks directly on resource-constrained embedded systems. To address this problem, we propose TaijiNet, a separable binary network, to reduce the storage and computational overhead while maintaining a comparable accuracy. Furthermore, we also introduce a strategy called partial binarized convolution which binarizes only unimportant kernels to efficiently balance network performance and accuracy. Our approach is evaluated on the CIFAR-10 and ImageNet datasets. The experimental results show that with the proposed TaijiNet, the separable binary versions of AlexNet and ResNet-18 can achieve 26x and 6.4x compression rates with comparable accuracy when comparing with the full-precision versions respectively. In addition, by adjusting the PCA threshold, the xnor version of Taiji-AlexNet improves accuracy by 4-8 percent comparing with other state-of-the-art methods.

Keywords:

Kernel

Convolution

Principal component analysis

Embedded systems

Analytical models

Quantization (signal)

Computational modeling

Convolutional neural network

binarization

embedded systems

Separable Binary Convolutional Neural Network on Embedded Systems

Abstract

AI Summary

Journal

Organization