吴正龙’s neural-networks Bookmarks
19 DEC 2016
[Andrej Karpathy] Yes you should understand backprop
Backpropagation is a leaky abstraction. It is a credit assignment scheme with non-trivial consequences. Ignoring how it works under the hood because "TensorFlow automagically makes my networks learn" means you won't be ready to wrestle with the dangers it presents.