I suddenly feel like implementing something
-
-
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
It was just an attempt to explain the relevant intuitions behind capturing spatial structure with neural networks for people that don't speak linear algebra, so don't expect depth or much novel insight :)
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
However, and going out on a limb, do you think we could train generative networks directly over parameterized shader functions, and use attention and metalearning instead of the chain rule for credit assignment?
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
The closing paragraph made me wonder: If we have prosthetic eyes/synthetic retinal stimulation (https://youtu.be/HOpNbAKAawE ), do we have actual footage of the raw visual input that leaves through our optical nerve (spatially re-assembled for us to see in 2d)?
-
Because I bet that would hammer the whole "the reality i see and experience is neither what is 'out there' nor 'just what enters my brain'"-point home. Tried looking around a bit but I can't find anything. Is sensor density the problem here?
- 3 more replies
New conversation -
-
-
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
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.