Deep learning is actually pretty boring, except for this one thing: it works. It's the real deal. Too bad it doesn't explain anything...
@fchollet (I really wish I understood why deep nets regularize so much better than they have any right to.)
-
-
@michael_nielsen@fchollet they don't? It's just that optimization sucks: finding the global optimum that overfits the train set is hard ;-) -
@syhw@michael_nielsen@fchollet that is: unless you're working with my current dataset (50 GB of data, overfits all the time!)
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.