Differences

This shows you the differences between two versions of the page.

--- data_mining:neural_network:tuning [2018/05/20 15:09] – [Algo] phreazer
+++ data_mining:neural_network:tuning [2018/05/20 15:21] (current) – [Batch norm at test time] phreazer
@@ Line 245: / Line 245: @@
   * For minibatch t:
-  *   Compute forward prop for $X^{\t{}}$
+    * Compute forward prop for $X^{\{t\}}$
-  *     In each hidden layer use BN to replace $Z^l$ with $\tilde{Z}^l$
+      * In each hidden layer use BN to replace $Z^l$ with $\tilde{Z}^l$
-  *   Use backprop to compute $dW^{l}, d\beta^{l}, d\gamma^{l}$
+  * Use backprop to compute $dW^{l}, d\beta^{l}, d\gamma^{l}$
-  *   Update parameters ...
+  * Update parameters ...
+==== Why does it work ====
+Covariance shift (shifting input distribution)
+  * Batch norm reduces amount in which hidden units shifts around, become more stable (input to later layers)
+  * Slight regularization effect: Adds some noise, because it's normed on the mini batch
+==== Batch norm at test time ====
+Here no mini-batch, but one sample at a time
+Estimate $\sigma^2, \mu$ using exponentially weighted average across mini-batches