1 of 14

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

2 of 14

Motivation��Old school related concept:��Feature scaling

  • The range of values of raw training data often varies widely
    • Example: Has kids feature in {0,1}
    • Value of car: $500-$100’sk
  • In machine learning algorithms, the functions involved in the optimization process are sensitive to normalization
    • For example: Distance between two points by the Euclidean distance. If one of the features has a broad range of values, the distance will be governed by this particular feature.
    • After, normalization, each feature contributes approximately proportionately to the final distance.
  • In general, Gradient descent converges much faster with feature scaling than without it.
  • Good practice for numerical stability for numerical calculations, and to avoid ill-conditioning when solving systems of equations.

​

3 of 14

Common normalizations

Two methods are usually used for rescaling or normalizing data:

​

  • Scaling data all numeric variables to the range [0,1]. One possible formula is given below:

​

​

​

  • To have zero mean and unit variance:

​

​

​

​

​

  • In the NN community this is call Whitening

4 of 14

Internal covariate shift:��The cup game example 

  •  The first guy tells the second guy, “go water the plants”, the second guy tells the third guy, “got water in your pants”, and so on until the last guy hears, “kite bang eat face monkey” or something totally wrong.

​

  • Let’s say that the problems are entirely systemic and due entirely to faulty red cups. Then, the situation is analogous to forward propagation

​

  • If can get new cups to fix the problem by trial and error, it would help to have a consistent way of passing messages in a more controlled and standardized (“normalized”) way. e.g: Same volume, same language, etc.�

“First layer parameters change and so the distribution of the input to your second layer changes”

5 of 14

Proposed Solution:�Batch Normalization (BN)

  • Batch Normalization (BN) is a normalization method/layer for neural networks.

​

  • Usually inputs to neural networks are normalized to either the range of [0, 1] or [-1, 1] or to mean=0 and variance=1

​

  • BN essentially performs Whitening to the intermediate layers of the networks.

​

6 of 14

Batch normalization �Why it is good?��

  • BN reduces Covariate Shift. That is the change in distribution of activation of a component. By using BN, each neuron's activation becomes (more or less) a Gaussian distribution, i.e. its usually not active, sometimes a bit active, rare very active.
  • Covariate Shift is undesirable, because the later layers have to keep adapting to the change of the type of distribution (instead of just to new distribution parameters, e.g. new mean and variance values for Gaussian distributions).
  • BN reduces effects of exploding and vanishing gradients, because every becomes roughly normal distributed. Without BN, low activations of one layer can lead to lower activations in the next layer, and then even lower ones in the next layer and so on.

​

7 of 14

The BN transformation is scalar invariant

  • Batch Normalization also makes training more resilient to the parameter scale.
  • Normally, large learning rates may increase the scale of layer parameters, which then amplify the gradient during backpropagation and lead to the model explosion
  • However, with Batch Normalization, backpropagation through a layer is unaffected by the scale of its parameters. Indeed, for a scalar a,

8 of 14

Batch normalization:�Other benefits in practice

  • BN reduces training times. (Because of less Covariate Shift, less exploding/vanishing gradients.)
  • BN reduces demand for regularization, e.g. dropout or L2 norm.

​

    • Because the means and variances are calculated over batches and therefore every normalized value depends on the current batch. I.e. the network can no longer just memorize values and their correct answers.)
  • BN allows higher learning rates. (Because of less danger of exploding/vanishing gradients.)
  • BN enables training with saturating nonlinearities in deep networks, e.g. sigmoid. (Because the normalization prevents them from getting stuck in saturating ranges, e.g. very high/low values for sigmoid.)

9 of 14

Batch normalization:�Better accuracy , faster.

BN applied to MNIST (a), and activations of a randomly selected neuron over time (b, c), where the middle line is the median activation, the top line is the 15th percentile and the bottom line is the 85th percentile.

�

10 of 14

Why the naïve approach Does not work?

  • Normalizes layer inputs to zero mean and unit variance. whitening.
  • Naive method: Train on a batch. Update model parameters. Then normalize. Doesn't work: Leads to exploding biases while distribution parameters (mean, variance) don't change.

​

    • If we do it this way gradient always ignores the effect that the normalization for the next batch would have
    • i.e. : “The issue with the above approach is that the gradient descent optimization does not take into account the fact that the normalization takes place”

​

11 of 14

The proposed solution: To add an extra regularization layer

A new layer is added so the gradient can “see” the normalization and make adjustments if needed.

12 of 14

Algorithm Summary:�Normalization via Mini-Batch Statistics�

​

  • Each feature (component) is normalized individually
  • Normalization according to:
    • componentNormalizedValue = (componentOldValue - E[component]) / sqrt(Var(component))
  • A new layer is added so the gradient can “see” the normalization and made adjustments if needed.
    • The new layer has the power to learn the identity function to de-normalize the features if necessary!
    • Full formula: newValue = gamma * componentNormalizedValue + beta (gamma and beta learned per component)
  • E and Var are estimated for each mini batch.
  • BN is fully differentiable.

13 of 14

The Batch Transformation: formally from the paper.�

14 of 14

Useful links

​

​

​