The h Bar

General chat for Physics SE (physics.stackexchange.com). For M...
Feb 13, 2020 19:17
D:
Feb 13, 2020 19:03
impostor!
Feb 11, 2020 17:43
@NovaliumCompany yeah, just need practice :)
Feb 10, 2020 22:35
go approximate away!
Feb 10, 2020 22:34
I would guess that if you just think of Neural Networks as powerful function approximators capable of modeling quite non-linear functions, you might be able to find some nice applications.
Feb 10, 2020 22:32
ymmv
Feb 10, 2020 22:31
I would probably try first by building an auto-encoder to do it :D
Feb 10, 2020 22:31
yeah..that's mostly a resolution up-scaling which is a different set of techniques I would guess
Feb 10, 2020 22:30
you should check it out
Feb 10, 2020 22:30
The results from Super Slo-mo are actually quite good :D
Feb 10, 2020 22:29
Neural networks are somewhat harder to "explain"
Feb 10, 2020 22:29
which can basically reduce them to a bunch of if-else statements
Feb 10, 2020 22:29
Trees also tend to tell you exactly how their nodes are split
Feb 10, 2020 22:28
@ACuriousMind actually explainable AI is much harder for neural-types of ML. Traditional ML algorithms tend to be surprisingly explainable... e.g. linear regression is pretty easy to explain
Feb 10, 2020 22:27
something about flows in there :)
Feb 10, 2020 22:27
@tpg2114 you ever see this paper on Super Slo-Mo? It uses optic flow to create interpolation frames to convert regular video into super slow-motion video :D
Feb 10, 2020 21:57
You need to deploy it to the right use case lol
Feb 10, 2020 21:57
well you probably shouldn't have a ML algorithm fly a plane by itself or something...
Feb 10, 2020 21:55
but there's a lot of rough edges that need to be straightened
Feb 10, 2020 21:55
It's pretty powerful
Feb 10, 2020 21:54
machine learning is a beautiful mess :)
Feb 10, 2020 21:37
seems legit
Feb 10, 2020 21:02
And then $e^{i\pi}$ is an integer...hmmm
Feb 10, 2020 21:00
no, $\pi$ doesn't know $e$
Feb 10, 2020 21:00
who knows
Feb 10, 2020 20:59
Someone should write a paper "On the Normaliciousness of Pi, E, and other Well Known Constants"
Feb 10, 2020 20:58
:D
Feb 10, 2020 20:58
Normaliciousness?
Feb 10, 2020 20:57
Normalcy?
Feb 10, 2020 20:57
Normal-ness is the property that describes this afaik
Feb 10, 2020 20:54
it hasn't been proven to be
Feb 10, 2020 20:54
I'm not sure, but I think that has to do with whether $\pi$ is a normal number or not
Feb 10, 2020 20:53
Also the Bible
Feb 10, 2020 20:53
so...some encoding of Moby Dick is in there somewhere
Feb 10, 2020 20:53
if it's a normal number then I think every possible sequence is somewhere in there?
Feb 10, 2020 20:52
for real
Feb 10, 2020 20:49
$\pi$ and stuff
Feb 10, 2020 19:03
cus companies will pull that shit on you lol
Feb 10, 2020 19:03
so you know...it's not like accepting then backing out is a "huge" no-no
Feb 10, 2020 19:02
word of caution: sometimes (not often, but sometimes) offers do get rescinded
Feb 10, 2020 19:01
Well, you can ask the recruiter when is your deadline for a response
Feb 10, 2020 19:00
well then you have to decide
Feb 10, 2020 18:59
but it does happen
Feb 10, 2020 18:59
Maybe not the most "polite" thing to do
Feb 10, 2020 18:59
You can still back out after accepting an offer
Feb 10, 2020 18:58
It does happen
Feb 10, 2020 18:58
Generally job offers don't last all that long..but I mean you could also accept the offer and then if you get a better offer soon, then tell them no.
Feb 10, 2020 18:22
Terminators gonna be like predators
Feb 10, 2020 18:22
Skynet is srsly becoming reality man...yesterday I saw another article about how the army is using ML on infrared imagery to do facial recognition in the dark
Feb 10, 2020 18:20
lol