It's possible that backprop has more success for narrower bounded problems and is "faster," but maybe because it's greedy. Biological learning might benefit fro

It's possible that backprop has more success for narrower bounded problems and is "faster," but maybe because it's greedy. Biological learning might benefit from being less greedy.