Sir, the link has different paper, not the paper written by Geoff Hinton.
Geoff Hinton says that deep networks are different from symbolic AI because they reason by analogy. Now we know how:https://arxiv.org/abs/2012.00152
-
-
-
No where did Pedro Domingos imply that the paper is from Hinton. What he wrote is that the paper answers the question by Hinton.
Keskustelun loppu
Uusi keskustelu -
-
-
Great synthesis, I would add this was somehow implied in the work of Deep Gaussian Process, and a recurrent topic on
@lawrennd talksKiitos. Käytämme tätä aikajanasi parantamiseen. KumoaKumoa
-
-
-
1/N It is an interesting exercise but useful? I am not sure. The whole dependence on C(t) makes it difficult (impossible?) to use this formulation to answer any of the interesting questions besides the fact that the whole formulation is about the total function and not
-
2/N the middle layers. It left you with questions like: What the heck, if it is a kernel machine why it is dependent on the order of training samples? Why the kernel itself depends on your choice of optimization algo? How one kernel is better than the other if they all are using
- Näytä vastaukset
Uusi keskustelu -
Lataaminen näyttää kestävän hetken.
Twitter saattaa olla ruuhkautunut tai ongelma on muuten hetkellinen. Yritä uudelleen tai käy Twitterin tilasivulla saadaksesi lisätietoja.