Mosaic-CNN: A Combined Two-Step Zero Prediction Approach to Trade off Accuracy and Computation Energy in Convolutional Neural Networks

Cheolhwan Kim, Dongyeob Shin, Bohun Kim, Jongsun Park

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

In convolutional neural networks (CNN), convolutional layers consume dominant portion of computation energy due to large amount of multiply-accumulate operations (MACs). However, those MACs become meaningless (zeroes) after rectified linear unit (ReLU) when the convolution results become negative. In this paper, we present an efficient approach to predict and skip the convolutions generating zero outputs. The proposed two-step zero prediction approach, called Mosaic CNN, can be effectively used for trading off classification accuracy for computation energy in CNN. In Mosaic CNN, the outputs of each convolutional layer are computed considering their spatial surroundings in an output feature map. Here, the types of spatial surroundings (mosaic types) can be selected to save computation energy at the expense of accuracy. In order to further save the computations, we also propose a most significant bits (MSBs) only computation scheme, where a constant value representing least significant bits (LSBs) compensates the MSBs only computations. The CNN accelerator supporting the combined two approaches has been implemented using 65nm CMOS process. The numerical results show that compared to the state-of-art processor, the proposed reconfigurable accelerator can achieve energy savings ranging from 16.99% to 29.64% for VGG-16 without seriously compromising the classification accuracy.

Original languageEnglish
JournalIEEE Journal on Emerging and Selected Topics in Circuits and Systems
DOIs
Publication statusAccepted/In press - 2018 Aug 11

Keywords

  • Circuits and systems
  • Computer architecture
  • Convolution
  • Convolutional neural networks
  • Convolutional neural networks
  • Energy efficiency
  • Energy-efficient accelerator
  • Kernel
  • Simulation

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Fingerprint Dive into the research topics of 'Mosaic-CNN: A Combined Two-Step Zero Prediction Approach to Trade off Accuracy and Computation Energy in Convolutional Neural Networks'. Together they form a unique fingerprint.

  • Cite this