1. Suppose the image is 5x5 and we do a 2x2 max pool with stride=2. In this case we have to pad the rightmost column and the bottom row with zeros. Padding also allows us to produce an output of the same dimensions as the input. For example below we pad the image before running a convolution 0 0 0 0 0 0 0 0 2 3 7 2 1 0 19 39 ? o ? 6 4 1 ? 5 0 2 4 2 0 ? ? ? 0 2 ? 4 2 3 1 4 5 8 9 2 0 56 721 ? ? ? ? 0 ? 0 1 0 0 2. ? ? ? ? 0 6 9 5 0 4 0 ? ? ? ? ? 0 0 0 0 0 0 0 nswer