Efficient batchwise dropout training using submatrices
Neural and Evolutionary Computing
2015-02-10 v1 Computer Vision and Pattern Recognition
Abstract
Dropout is a popular technique for regularizing artificial neural networks. Dropout networks are generally trained by minibatch gradient descent with a dropout mask turning off some of the units---a different pattern of dropout is applied to every sample in the minibatch. We explore a very simple alternative to the dropout mask. Instead of masking dropped out units by setting them to zero, we perform matrix multiplication using a submatrix of the weight matrix---unneeded hidden units are never calculated. Performing dropout batchwise, so that one pattern of dropout is used for each sample in a minibatch, we can substantially reduce training times. Batchwise dropout can be used with fully-connected and convolutional neural networks.
Cite
@article{arxiv.1502.02478,
title = {Efficient batchwise dropout training using submatrices},
author = {Ben Graham and Jeremy Reizenstein and Leigh Robinson},
journal= {arXiv preprint arXiv:1502.02478},
year = {2015}
}