It’s a basic ML fact that there’s an infinity of laws compatible with the same observations. Which ones we believe is more a function of psychology than physics.
-
Reality itself is symmetric. Yes, you can use that as an inductive bias, but the symmetries in reality came before the symmetries as an inductive bias.
-
-
I think it's actually warranted to use stronger language than "assume"? We can reflect on where our inductive biases came from, and be like, "Ah, the symmetries in physics are such that evolution by natural selection would favor brains with corresponding priors"
-
It's also extremely noteworthy that the symmetries appear to hold. That is, you can perfectly predict the motions of objects using only laws that are symmetric to translations, boosts, etc..
- 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.