Human brains are a product of genetic optimisation and deep learning. Genetic optimisation accounts for the local minima problem in gradient descent by exploring the vicinity of the manifold. The more we know about the manifold the better our odds of finding global minima.
Deep learning teaches us something about evolution: if it's easy to find the global optimum in a high-dimensional space, maybe we're not so random after all.
-
-
Kiitos. Käytämme tätä aikajanasi parantamiseen. KumoaKumoa
-
-
-
Evolution is random search, I guess.
Kiitos. Käytämme tätä aikajanasi parantamiseen. KumoaKumoa
-
-
-
Pedro write more books please
-
oh god no
Keskustelun loppu
Uusi keskustelu -
-
-
It isn't easy. There's plenty of examples in species where evolution has reached lowsy designs. And we cannot say we are the optimum result of it. On what grounds?
-
Optimal for each niche.
Keskustelun loppu
Uusi keskustelu -
-
-
No,deep learning has nothing to teach about evolution,but it does teach us a lot about having deep pockets... it can delude you into thinking that deep learning is somehow intelligent, or worse that knowing anything about it,makes one artificially intelligent. Spoiler: It doesn't
-
deep learning needs deep pockets..
Keskustelun loppu
Uusi keskustelu -
-
-
Evolution is anything but random. And I don't think there's any definition of evolution that asserts it's a random process. By definition, organisms are shaped by environmental pressures aka natural selection.
-
Evolution requires random mutations.
- 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.