English

meProp: Sparsified Back Propagation for Accelerated Deep Learning with Reduced Overfitting

Machine Learning 2019-03-12 v5 Artificial Intelligence Computation and Language Computer Vision and Pattern Recognition

Abstract

We propose a simple yet effective technique for neural network learning. The forward propagation is computed as usual. In back propagation, only a small subset of the full gradient is computed to update the model parameters. The gradient vectors are sparsified in such a way that only the top-kk elements (in terms of magnitude) are kept. As a result, only kk rows or columns (depending on the layout) of the weight matrix are modified, leading to a linear reduction (kk divided by the vector dimension) in the computational cost. Surprisingly, experimental results demonstrate that we can update only 1-4% of the weights at each back propagation pass. This does not result in a larger number of training iterations. More interestingly, the accuracy of the resulting models is actually improved rather than degraded, and a detailed analysis is given. The code is available at https://github.com/lancopku/meProp

Keywords

Cite

@article{arxiv.1706.06197,
  title  = {meProp: Sparsified Back Propagation for Accelerated Deep Learning with Reduced Overfitting},
  author = {Xu Sun and Xuancheng Ren and Shuming Ma and Houfeng Wang},
  journal= {arXiv preprint arXiv:1706.06197},
  year   = {2019}
}

Comments

Accepted by the 34th International Conference on Machine Learning (ICML 2017)

R2 v1 2026-06-22T20:23:21.371Z