BFConv: Improving Convolutional Neural Networks with Butterfly Convolution

  • Dengjie Yang
  • , Xuehui Yu
  • , Yi Sun
  • , Fuzhen Zhuang
  • , Qing He
  • , Shiwei Ye*
  • *Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Convolutional neural network (CNN) is a basic neural network widely used in vision tasks. Many CNNs alleviate the redundancy in feature maps to reduce model complexity. Inspired by digital signal processing theories, this paper reviews discrete fourier transform (DFT), finding its similarities with standard convolution. In particular, DFT has a fast algorithm called FFT, which sparks our thinking: can we learn from the idea of FFT to realize a more efficient convolution filter? Based on the butterfly operation of FFT, we propose a novel butterfly convolution (BFConv). In addition, we illustrate that group weight sharing convolution is a basic unit of BFConv. Compared with the traditional group convolution structure, BFConv constructs group residual-like connections and increases the range of receptive fields for each sub-feature layer. Without changing the network architecture, we integrate BFConv into ResNet-50, ShuffleNet and VGG-16. Experimental results on CIFAR-10 and ImageNet demonstrate the above BFConv-equipped networks reduce parameters and computation, achieving similar or higher accuracy. Remarkably, when ResNet-50 embedded BFConv reaches nearly half of the compression ratio of the model, it performs favorably against its state-of-the-art competitors.

Original languageEnglish
Title of host publicationNeural Information Processing - 28th International Conference, ICONIP 2021, Proceedings
EditorsTeddy Mantoro, Minho Lee, Media Anugerah Ayu, Kok Wai Wong, Achmad Nizar Hidayanto
PublisherSpringer Science and Business Media Deutschland GmbH
Pages40-50
Number of pages11
ISBN (Print)9783030922726
DOIs
StatePublished - 2021
Externally publishedYes
Event28th International Conference on Neural Information Processing, ICONIP 2021 - Virtual, Online
Duration: 8 Dec 202112 Dec 2021

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume13111 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference28th International Conference on Neural Information Processing, ICONIP 2021
CityVirtual, Online
Period8/12/2112/12/21

Keywords

  • Butterfly convolution
  • Convolutional neural network
  • FFT
  • Group Weight sharing convolution

Fingerprint

Dive into the research topics of 'BFConv: Improving Convolutional Neural Networks with Butterfly Convolution'. Together they form a unique fingerprint.

Cite this