[hnarayanan]

I don't know anything about anything but I'm willing to bet that the end result is very similar. They're "just" using neural networks as rich approximators.