1 paper · 1 filter
Henry Z. Lo, Kevin Amaral, Wei Ding
One of the difficulties of training deep neural networks is caused by improper scaling between layers. Scaling issues introduce exploding / gradient problems, and have typically be…