Why this matters: If you have a generator of interesting artifacts, and you let a user interact with it, they should be able to get "momentum", moving from a bad thing to a good thing, to a better thing. But this requires "smoothness" in the possibility space.
-
Show this thread
-
Replying to @GalaxyKate
so: a change in the input space creates a perceptibly proportionate change in the output space? I'm no mathematician, but in neuroscience these are "vector spaces" or "smooth representational spaces" (
@o_guest has written a nice paper about smooth representations in the brain)1 reply 0 retweets 2 likes -
the section on "functional smoothness" describes something like what you are describing https://elifesciences.org/articles/21397
1 reply 0 retweets 2 likes -
although in the case you describe it seems important that the mappings are on the scale of the human perceptual system, which is less a math thing
1 reply 0 retweets 1 like -
Replying to @beausievers @o_guest
Its half human-perception/half-math? Like I can model "prettiness of flowers parametrically generated from N-dimensional vectors" which is squishy human stuff, but also "can I cluster those flowers with ML or lerp smoothly between them" which is less so
1 reply 0 retweets 2 likes -
SQUISHY MATH
1 reply 0 retweets 2 likes -
Replying to @GalaxyKate @o_guest
and I am extremely not a mathematician, so I could be wrong about this, but (setting human perception aside), you may be looking for "smooth maps between manifolds"—(but i am not a mathematician)
1 reply 0 retweets 1 like -
Replying to @beausievers @o_guest
I am also not a mathematician but balancing how much to learn/pretend in order to get useful things done
1 reply 0 retweets 2 likes -
which is probably a working definition of a mathematician
1 reply 0 retweets 2 likes -
Replying to @GalaxyKate @beausievers
Thanks for the link to our paper — yes, what
@ProfData and I wrote about deas seem to (at least superficially) match what you describe. For example, it's possible to get both subjective measures of similarity (ask people what they think) and "objective" measures using fMRI voxels1 reply 0 retweets 1 like
and not only are they comparable they are also often similar in important ways. My slides might shed some light on this if you are interested/if it's relevant: https://figshare.com/articles/What_the_Success_of_Brain_Imaging_Implies_about_the_Neural_Code/4252022 …
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.