Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift ​

← Back to topic

Authors: Sergey Ioffe, Christian Szegedy
Year: 2015
Journal: ICML
DOI: 10.48550/arXiv.1502.03167
Publisher: https://arxiv.org/abs/1502.03167

Keywords: batch normalization, deep learning

Abstract ​

Training Deep Neural Networks is complicated by the fact that the distribution of each layer's inputs changes during training as the parameters of the previous layers change.

Cite this paper ​

@misc{bnicml2015,
  title  = {Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift},
  author = {Sergey Ioffe, Christian Szegedy},
  year   = {2015},
  journal = {ICML},
  doi    = {10.48550/arXiv.1502.03167},
  url    = {https://doi.org/10.48550/arXiv.1502.03167},
}

Source files ​

• metadata.json
• paper.bib