I’m looking forward to the concrete out-of-sample predictions that will be made by any deep learning theory. So far, they’ve been mostly ex-post rationalizations of properties that have already been observed in practice, which are quite easy to produce at an informal level.
I’m looking forward to the concrete out-of-sample predictions that will be made by any deep learning theory. So far, they’ve been mostly ex-post rationalizations of properties that have already been observed in practice, which are quite easy to produce at an informal level.