pith. sign in

arxiv: 1805.11604 · v5 · pith:HXZA4IWWnew · submitted 2018-05-29 · 📊 stat.ML · cs.LG· cs.NE

How Does Batch Normalization Help Optimization?

classification 📊 stat.ML cs.LGcs.NE
keywords batchnormtrainingbatcheffectivenessfasternormalizationoptimizationstable
0
0 comments X
read the original abstract

Batch Normalization (BatchNorm) is a widely adopted technique that enables faster and more stable training of deep neural networks (DNNs). Despite its pervasiveness, the exact reasons for BatchNorm's effectiveness are still poorly understood. The popular belief is that this effectiveness stems from controlling the change of the layers' input distributions during training to reduce the so-called "internal covariate shift". In this work, we demonstrate that such distributional stability of layer inputs has little to do with the success of BatchNorm. Instead, we uncover a more fundamental impact of BatchNorm on the training process: it makes the optimization landscape significantly smoother. This smoothness induces a more predictive and stable behavior of the gradients, allowing for faster training.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 4 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Geometric Deep Learning: Grids, Groups, Graphs, Geodesics, and Gauges

    cs.LG 2021-04 accept novelty 6.0

    Geometric deep learning provides a unified mathematical framework based on grids, groups, graphs, geodesics, and gauges to explain and extend neural network architectures by incorporating physical regularities.

  2. Gradient Noise Convolution (GNC): Smoothing Loss Function for Distributed Large-Batch SGD

    cs.LG 2019-06 unverdicted novelty 5.0

    GNC convolves stochastic gradient noise to smooth sharp minima in large-batch SGD, outperforming isotropic noise for better generalization in distributed deep learning.

  3. Autoencoding sensory substitution

    q-bio.NC 2019-07 unverdicted novelty 4.0

    Deep recurrent autoencoders convert images to shortened audio signals that incorporate hearing models, enabling above-chance hand posture discrimination and object reaching after a few hours of training instead of months.

  4. Mean Spectral Normalization of Deep Neural Networks for Embedded Automation

    cs.LG 2019-07 unverdicted novelty 4.0

    Proposes MSN reparameterization to address mean-drift in SN, claiming ~16% faster inference than BN with fewer parameters on CNNs and GANs.