Batch renormalization in Deep Learning - June 2017
An interesting article about Batch renomarlization for Deep Learning.
Authors prove the efficiency of Batch Renormalization for Deep Learning.
Deep Learning has revolutionized vision via convolutional neural networks (CNNs)
and natural language processing via recurrent neural networks (RNNs). However,
success stories of Deep Learning with standard feed-forward neural networks
(FNNs) are rare. FNNs that perform well are typically shallow and, therefore cannot
exploit many levels of abstract representations. We introduce self-normalizing
neural networks (SNNs) to enable high-level abstract representations. While
batch normalization requires explicit normalization, neuron activations of SNNs
automatically converge towards zero mean and unit variance. The activation
function of SNNs are “scaled exponential linear units” (SELUs), which induce
self-normalizing properties. Using the Banach fixed-point theorem, we prove that
activations close to zero mean and unit variance that are propagated through many
network layers will converge towards zero mean and unit variance — even under
the presence of noise and perturbations.
For reading, this is here:
https://arxiv.org/pdf/1706.02515.pdf