Weight Normalization: A Simple Reparameterization to Accelerate Training of Deep Neural Networks

@inproceedings{Salimans2016WeightNA,
title={Weight Normalization: A Simple Reparameterization to Accelerate Training of Deep Neural Networks},
author={Tim Salimans and Diederik P. Kingma},
booktitle={NIPS},
year={2016}
}

We present weight normalization: a reparameterization of the weight vectors in a neural network that decouples the length of those weight vectors from their direction. By reparameterizing the weights in this way we improve the conditioning of the optimization problem and we speed up convergence of stochastic gradient descent. Our reparameterization is inspired by batch normalization but does not introduce any dependencies between the examples in a minibatch. This means that our method can alsoâ€¦Â CONTINUE READING