English

Spatial Correlation and Value Prediction in Convolutional Neural Networks

Computer Vision and Pattern Recognition 2019-01-03 v2

Abstract

Convolutional neural networks (CNNs) are a widely used form of deep neural networks, introducing state-of-the-art results for different problems such as image classification, computer vision tasks, and speech recognition. However, CNNs are compute intensive, requiring billions of multiply-accumulate (MAC) operations per input. To reduce the number of MACs in CNNs, we propose a value prediction method that exploits the spatial correlation of zero-valued activations within the CNN output feature maps, thereby saving convolution operations. Our method reduces the number of MAC operations by 30.4%, averaged on three modern CNNs for ImageNet, with top-1 accuracy degradation of 1.7%, and top-5 accuracy degradation of 1.1%.

Keywords

Cite

@article{arxiv.1807.10598,
  title  = {Spatial Correlation and Value Prediction in Convolutional Neural Networks},
  author = {Gil Shomron and Uri Weiser},
  journal= {arXiv preprint arXiv:1807.10598},
  year   = {2019}
}

Comments

This paper has been accepted to IEEE Computer Architecture Letters (https://ieeexplore.ieee.org/document/8594568)

R2 v1 2026-06-23T03:16:56.659Z