The race in AI should be toward models with fewer and fewer parameters for a given performance, not the opposite.
-
Yes that's exactly what I meant: nothing wrong with "generalization" (aka accuracy on unseen data) but the "textbook statistical learning" (aka assuming same distribution for training and test) might not be the right lens for deep learning.
-
-
Not at all. What I'm trying to say is that "smoothness is not required for generalization" is itself a questionable statement because a) we don't have the right definition of generalization for DL while b) at least smoothness is well-defined and "feels" like baby generalization.
-
There are plenty of cases of good generalization that don’t require smoothness. You’re overfitting to (current) DL :)
Keskustelun loppu
Uusi keskustelu -
Lataaminen näyttää kestävän hetken.
Twitter saattaa olla ruuhkautunut tai ongelma on muuten hetkellinen. Yritä uudelleen tai käy Twitterin tilasivulla saadaksesi lisätietoja.