>images created by a neural net.
The era of synthetic porn is upon us.
>>54481707
Neat.
>>54481707
>thinking this exists
>believing in the charade
The end is nigh, stop being distracted by ``pretty'' pictures.
>>54481816
gb2 /pol/
>>54481707
>images created by a neural net
>by mashing other images of the same thing together
I don't come to /g/ much but it had the best ArfaGo threads and I had a good impression.
What do you think about something like this >>>/sci/8062657 here? Would a /ml/ have any hopes of survival here? The idea of operating on content in 4chan makes me really happy.
As you can see from the /sci/ thread, there seems to be very little interest, perhaps too busy with their tropical algebras on banana spaces.
Maybe /g/ is too fast though, I don't know.
>>54481844
Idk,, that sounds pretty /g/ to me
>>54481884
> mashing other (...) of the same (... Implying gender) together
There you go ... Porn!
>>54482236
I have tried discussing /ml/ here a few times and responses vary. Although I am no expert myself I would love to learn more.
For the most part I don't think a whole lot of people here are too familiar with anything other than dicking around with their PCs, but I could be wrong.
>>54481816
Why do you use `` and '' instead of "?
>>54482404
I'm not an expert either, but I went through Ng's coursera. I think it's actually a lot easier to get into than most people expect.
And anyhow, I think people being interested is more important than having experts on it around, though I wouldn't be surprised if occasionally some people pop in with experience in it.
I haven't lurked /g/ enough to see how fast it moves, keeping a low-interest general alive can be painful. (the /tg/ go/baduk thread after the AG stuff was well-intentioned, but ultimately a failure)
>>54481707
is this what ive been doing all those retarded picture captchas for?
i-it's... us
>>54482479
Give it a shot, you never know. It would be a higher quality thread than most and actually on topic but those are the ones that not many people seem to jump into.
>>54482529
but shit this is already page 3. How can it survive?
It really does sound fun though. We could do analytics on the thread itself, and /g/, to see average post-rate needed to stay alive, pointless stuff like that. (I guess I'm imagining a combination data science/ML general)
>>54482404
>>54482479
>>54482636
Tried many many times to make a machine learning general on /g/, but it never survives.
/g/ just doesn't care about ai.
>>54482636
This is a shitty thread that doesn't prompt discussion. Since most people are not actively doing any ML projects, even if they are interested, they won't have much to say. By getting some good links or good examples the thread might have a shot.
Love the idea of applying some concepts to the thread itself, though. Exactly how I think but I never have the skill when I have the curiosity and vice versa.
>>54481707
So it's basically a distorted, reduced version of whatever has been used to train the neural network. Doesn't seem too interesting.
>>54482666
Shitty news but kind of expected. I saw one of those generals and it was completely overshadowed by a 6 word drakepost on some AMD vs Nvidia shit. Like 13 responses Vs 200 in the drakepost. Maybe I really should just go to reddit
>>54482666
Dang but clearly a few of us do. It would feel shitty to start one on /sci/ and then shill for it on /g/.
>>54482702
So a good OP might go far in establishing a community of /ml/ers. /ML/? /pat/? /mach/?
Let's post some links for potential OP, and keep brainstorming on how to keep a ML/(and data science?) general alive
Intro-tier:
>https://medium.com/@ageitgey/machine-learning-is-fun-80ea3ec3c471#.6wkt9refp
>https://www.youtube.com/watch?v=bxe2T-V8XRs&list=PLiaHhY2iBX9hdHaRr6b7XevZtgZRa1PoU
>>54482762
I'd love for a ML / data science general on /g/. I've just recently started in the field (last October), but I'd love to share whatever little I know with some gentoomen hier.
>>54482762
Essential: https://github.com/4chan/4chan-API
Pretty easy to use the API. Moderately easy to get into ML.
Could be a fun general. I really like the idea of surprising some regular in /ic/ with art generated in their own style. Or gobbling up /lit/ for a couple weeks and seeing what each of our algorithms spits out as a /lit/ post.
Or doing a kaggle competition and winning a million bajillion dollars that we donate to the EFF
/mlds/ - Machine Learning and Data Science
>Beginner links
https://medium.com/@ageitgey/machine-learning-is-fun-80ea3ec3c471#.6wkt9refp
https://www.youtube.com/watch?v=bxe2T-V8XRs
>/mlds/ projects (?)
Course by Andrew Ng, on the theoretical side, uses matlab/octave
https://www.coursera.org/learn/machine-learning/
Another ML course, uses R
https://www.coursera.org/learn/practical-machine-learning/
https://lagunita.stanford.edu/courses/HumanitiesSciences/StatLearning/Winter2016/about
>2 free books by the instructors of above course:
http://www-bcf.usc.edu/~gareth/ISL/
>datasets (?)
probably too big of a category because there's a fuckton of available datasets out there, probably only put compilations of datasets here
>libraries (?)
>>54484346
>May project (or Project 1, or ?)
(something about /g/ or /mlds/)
>4chan API
https://github.com/4chan/4chan-API
>Interesting projects, or other ML stuff, or ?
https://www.jukedeck.com
http://cs229.stanford.edu/projects2013.html - for ideas on projects, change year in URL for more
http://imgur.com/a/K4RWn - challenge images
http://vizdoom.cs.put.edu.pl/competition-cig-2016 - visual doom AI competition
https://www.kaggle.com/ - ML competitions with rewards
https://github.com/alexjc/neural-doodle
https://github.com/awentzonline/image-analogies
>Cheatsheets, infographics
http://www.datasciencecentral.com/profiles/blogs/20-data-science-r-python-excel-and-machine-learning-cheat-sheets
Comments, organization, show of interest all appreciated
WE CAN DO THIS
http://pastebin.com/5a3E6Y5g a bit of my shitty code for scraping (in R)
Anyone have any ideas for potential machine learning or data analytic type things to perform on 4chan?
Definitely interested in seeing this general go through. I'm a complete newb with ML.
>>54486209
cool, same here on both accounts. I'm thinking this could be the push to get me going on it. The shit I've seen with ML is pretty nice, and I hear it's not as crazy math-heavy as it seems. I think it can be but also libraries.
also I don't care if you fucks think I'm an idiot for asking something so that helps.
>>54486389
>pic related
>>54484346
>>54485211
Feedback on any of this? I'm peeing in my pants excited about an ML/DS general. It's also lucrative as fuck, and is only getting easier to implement as companies are getting thirstier for it. Really pumped that others are into this too. Still kind of keeping an eye on page rate as some sort of indicator of how active it would need to be to survive (gee wouldn't it be nice to source this to a machine...)
Seems like after a bump things quickly move to page 3-5 and then slow down. I suppose there are a couple other generals here that people post in regularly, I haven't been watching the catalog and I haven't been to /g/ since alphago stuff.
Also I love machine learning because it's not a "set it up and let it run" technique, it requires some finesse and intuition. And let me remind you:
https://www.kaggle.com/competitions
$$$
Is neural network CP illegal or legal?
https://www.youtube.com/watch?v=S_f2qV2_U00
Not too fond of the forced haha, but it's good and the results speak for themselves.
this drops to page ~4 then sit steady and slowly drops?
Hypothesis? might involve average length of new threads. Finding features is a big part of machine learning.
This is the problem with the world today. Millennials think they know everything and that computers can comprehend things such as god or beauty.
>>54486925
Time for thoughtcrimes to actually be a thing.
>>54487552
>implying your personal imagination of a god actually exists outside your mind
>>54486654
Do you remember the website for the types in your image?
>>54482425
``autism´´
>>54487671
if I don't, do you think we could make something like this?
TensorFlow, BigQuery, what are some other big industry folks interested in Machine Learning?
Are you excited about this? I'm excited about this. Looking for some comment on a useful OP.
>>54484346
>>54485211
Also 4chan micro-communities we could perform data analytics on.
https://www.youtube.com/watch?v=xeYfK9NQyoo
>>54482666
would work better if we still had a text-only board
have a bump, op
back when pol tried to make their own tay-ai it seemed like there was a lot of interest in doing ml/nn stuff but as with all 4chan projects it died
so yes, no one actually knows how to do anything on this board other than rice their shitty desktops
>>54487792
>no one actually knows how to do anything on this board other than rice their shitty desktops
But yes!! This is excellent. Nobody has to give a fuck about asking stupid questions, or showing off their shitty code. It's perfect m8.
WHAT SHOULD WE DO, HOW SHOULD WE DO IT
TensorFlow, BigQuery, ....
This is BIG SHIT. Apparently only getting bigger, if my math degree and ear-to-the-floor serves. So far it's "applied statistics" but there's more there, a lot more. This involves procedural generation^newlevel, consider NPC behavior built on player behavior
$$$ stupid to ignore this
