Efficient neural network using pointwise convolution kernels with linear phase constraint

Journal Article (Journal Article)

In current efficient convolutional neural networks, 1 × 1 convolution is widely used. However, the amount of computation and the number of parameters of 1 × 1 convolution layers account for a large part of these neural network models. In this paper, we propose to use linear-phase pointwise convolution kernels (LPPC kernels) to reduce the computational complexities and storage costs of these neural networks. We design four types of LPPC kernels based on the parity of the number of input channels and symmetry of the weights of the pointwise convolution kernel. Experimental results show that Type-I LPPC kernels can compress some popular networks better with a small reduction in accuracy than the other types of LPPC kernels. The LPPC kernels can be used as new 1 × 1 convolution kernels to design efficient neural network architectures in the future. Moreover, the LPPC kernels are friendly to low-power hardware accelerator design to achieve lower memory access cost and smaller model size.

Full Text

Duke Authors

Cited Authors

  • Liang, F; Tian, Z; Dong, M; Cheng, S; Sun, L; Li, H; Chen, Y; Zhang, G

Published Date

  • January 29, 2021

Published In

Volume / Issue

  • 423 /

Start / End Page

  • 572 - 579

Electronic International Standard Serial Number (EISSN)

  • 1872-8286

International Standard Serial Number (ISSN)

  • 0925-2312

Digital Object Identifier (DOI)

  • 10.1016/j.neucom.2020.10.067

Citation Source

  • Scopus