I’ve always felt this way because the brain simply mixes things up from what it has got as sensory inputs. This is probably why you can’t imagine a color that you never saw or imagine a audio tone which you’ve never heard. Is there a talk of yours on this I can listen to?
-
-
-
Our recent work builds recurrent feedback into neural networks in form of a generative model: https://arxiv.org/abs/2007.09200
@YujiaHuangC@JamesGornet@TanNguyen689@ZhidingYu@doristsao - Show replies
New conversation -
-
-
I'd say that if you can't imagine one, I guess you don't. :)
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Without generative model it would ne hard to imagine.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
My bet is most almost certainly (with neurobiological evidence being dreams imagination etc...) But what we do know for certain is our brains generated generative models in ML :)
-
haha, so brain a meta-generative model? :-) Our recent preprint places a lot of neurobiological evidence in the context of inference in a generative model:https://www.biorxiv.org/content/10.1101/2020.09.09.290601v1 …
End of conversation
New conversation -
-
-
Probably yes. Dreams could be the result of generative models in the brain running unchecked by any sensory input.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Absolutely! We do have an internal generative model that gets updated with the sensory data to bridge the gap between it and the real world.
-
Yes - a lot of current theory built around the idea that most of what we "experience" is actually generated, and adjusted to various degrees by sensory-motor input/output. Surprise occurs when things do not conform well to our generative model's expectations.
End of conversation
New conversation -
-
-
Great question. It feels that way when meditating. Also, our eyes saccade and have blind spots that we are not aware of. Is what we see largely generated? Or is it like rapidly moving TV frames giving us a sense of full view? Regardless, generation is a good hypothesis for AI.
-
Aren't we just smart StyleGAN when drawing an art?
- Show replies
New conversation -
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.