TINDER ENDGAME: EUGENICS

Hello anons. After getting tired of swiping and getting latinos and all sorts of mutty filth, I decided that technology today is good enough that we can solve this problem manually. Today, I put the finishing touches on a general adverserial neural network ("GAN"), a type of neural network ("AI"), that is particularly good at identifying traits in images. Basically what you do is have a bunch of tagged data of traits (e.g. blond hair, weight, height, lips, race, make-up, and 65 other markers) meaning it can detect 71 traits in a person with its current set of training data.

The network it uses was actually a learning experience to implement a GAN from scratch in C#, but then I decided to evolve it. It automagically swipes, and saves, any girls that match your preference. Tinder has a bunch of interesting data for each profile, including music, Instagram account, etc. that can be used for further investigation. It has a pretty good accuracy. I went through a sample size it had picked out from me of NOPE/YES, and it had an 96% accuracy (144 correctly classified out of 150 total). All were identified correctly as being white, there was just some makeup flaws that it misdetected 6 times. I do not like make-up, especially not excessive make-up.

I'm thinking of making this a SaaS, since it's too GPU-heavy to run in the browser, I cannot make it an open-sourced extension. I can make it an extension that contacts my GPU-farm and does the work for anons. I wonder if this is something Jow Forums would be interested in.

I call the projects Tinder E-ugnics.
My personal background is a masters in Computer Science working towards a PhD.

Attached: TinderEugenics.png (1190x727, 650K)

Other urls found in this thread:

docs.opencv.org/2.4/modules/ml/doc/neural_networks.html
twitter.com/NSFWRedditImage

>is this smart
>wastes his time tinkering with (((tinder))) rather than making white babies
Best goy

Bump

>Is such a narrow minded faggot he can't appreciate based Jow Forums turbo autism

Based danebro.

>My personal background is a masters in Computer Science working towards a PhD.

Now go hack Pentagon. If a silly britbong rapebaby can do it so can you.

I do appreciate it. The first thing I said was he is smart, but frivolously wasting your intelligence without passing it on is disgenic

I love it Danskvän!

Also: Why are non-white girls the most thirsty on tinder?

sounds interesting

Sounds brilliant, and I would assume it could be adapted across a multitude of platforms?

BASED AND REDPILLED

>one post by this ID
Post more than a screenshot pls

Can you create an algorithm that identifies coalburners and whores who have slept with over 5 different men?

I’m getting into deep learning, but in a completely different field that can actually make money (financial markets)

Can you run these things on a regular computer at home with a

I like this idea!

How big is your dataset?

Very interesting project, Danebro!
The only thing I find regrettable is that you probably won't get university funding to work on it.

Attached: 1552550607654.jpg (400x399, 24K)

Wont financial markets be rendered fucking useless if it’s all rigged by high-frequency trading? Really what the fuck is the point in investing in an artificial market with fake investors?

Absolute genius user

I am this Ai

>TINDER ENDGAME: EUGENICS Geneticist
where is your github fine sir ?

you have the most greatest idea to be discrectionary and discriminatory at the same time

i would love to download this shit asap
or use it for linkedin surveys

If you jailbreak an Iphone you can bypass the paywall and put an auto swipe on. Then you just look in an hour and unmatch the fuglys

offer this to women
women usually are to lazy to swipe and guys usually always swipe right on them
tinder makes up for this and shows women only the hottest guys in their area
so their matches stack up quickly, but chad doesn't answers and beta boy doesn't get any matches regardless how much he swipes

OK, I'd like a 175cm red haired (copper colour) gf with green eyes, please.

you should make this open source
i am thinking about stealing you idea dane bro

this, OP's target audience certainly isn't Jow Forums

did you manually label the data?
i've been thinking of doing something similar but can't be arsed to label data.

Make it able to detect how many sexual partners she has had and I will gladly pay for your SaaS.

Same here we get along i am the neural part

thanks
>pic related

Attached: looks_like_she_found_one.png (720x1280, 868K)

Can you make this work outside of tinder? Towards where it can scan all pictures

>detect how many sexual partners she has had
Its tinder user

mad lad

actually....how do we get this to every individual on Jow Forums that needs to do hiring for large companies. I'm pretty sure you can run this without anyone knowing

Attached: 1555255179443.gif (480x270, 1.34M)

>never go against the trend
Long term stocks as well.

You best put in that request with Geppetto

As a rough guide I mean, you get roasties and then you get ultra-roasties. Apparently some traits make women more promiscuous such as a pronounced jawline so that could be incorporated.

Yes it will once the billion dollar high frequency trading firms get their hands on the best AI to make the market extremely efficient. Then not only do you need the most intelligent AI to make money in the market, you also need the speed (faster connections to the exchange through billion dollar underground cables) which very few people are going to have access to

That's why it would be so cool if there was university funding for it.

What we used to do with our training data is get an 8 euro/hour lower semester student to annotate the data. The chair would finance that and students would be happy to make some extra money with easy work and the convenience of doing it right at the university.

We have done something like that with cuneiform writing on Hittite tablets and with labelling windows on building fronts for a vertical solar potential catastre.

obviously he is trying to increase his chances of passing his genes on through a fitter female and not to just dump his semen into the first whore faking the best behavior to trick him
see the broader picture user

Its actually quite easy user i never would have thought of this idea, until one of the greatest dane's who ever browsed this Mongolian web weaving forum came up with it.
But honestly i think it should be open source i would pay for it but sharing is caring..

good work, user, glad to see we have such smart Jow Forumsacks

what number would you be comfortable with?

Roughly 6400 individual profiles (thank god for Tinder Gold), with an average of 2.34 images per profile. What you do to make sure the network can identify traits from different angles etc. is transform the images, e.g. flip, rotate, roll, stretch, that generates many times the amount of images you had originally (up to 10x as many!). That's why it's so very accurate.

Even though my background is in computer science, I have read a lot about genetics and derived physical and mental traits in the creation of this project. I am a big proponent of race homogeneity. Denmark is a relatively peaceful and homogeneous (so far), and I know people of other phenotypes may use this to prey on white chicks.

Attached: Eugenics_Project_4chan_Preview.png (1434x615, 1.63M)

You're a faggot, this guy has been programming Jow Forums tier shit for years here.

Based.

The AI is invading Jow Forums and the ideas among spergs are fucking brilliant.

I also have a lot of MPI/C#/C++ experience, but I want to use it for the sake of eliminating incels (not killing, of course) and balancing the gap 80/20 (80% of wamen fuck 20% of Chads) -- obviously, in reality the figures are even worse.

it'd be possible, and relatively easy, to just implement face detection algorithms and check the pixelcolor content of the detected faces.
then the user can set some thresholds for automated swiping.

however, in reality this would be part of creating some stupid rube goldberg machine to solve a problem which wouldn't exist if a proper dating is used or developed.

Next minister of Racial Relations of the revived Kalmar Union has been found!
(2022, Copenhagen, Denmark)

I have to go now for roughly 30 minutes. Going home from University (that's why I tripfagged, so I can verify on another IP).

If you have any inquiries message me at [email protected]

Glad you all appreciate the idea. I've read the responses, I realize its capabilities. This is not only specific to Tinder. I will have to find a way to share it without undesirables getting their hands on it.

See you later Jow Forums. Keep the thread going.

Unironicly based. Today OP was a not a faggot. Bump for interest. I suggest doing a GitHub dump if possible. To spread this around. Keep up the good work Danebro.

Enjoy getting arrested by the diversity glowniggers

Absolute Legend! how long until we can use your technology?

i thought about doing this with openCV
glad someone else did it

so yes, I'm interested
and if you need someone who knows Java enough to be barely useless, I'm your American

>coalburners and whores who have slept with over 5 men
>slept with over 5 men
that would eliminate like 99.999% of tinder users, you'd be left with undesirables
(yes, users. many faggots go on there kek)

so this is the power of autism

Can't wait for all the thots and (((media))) to come out against this.

Attached: 1554169894016.jpg (1134x597, 339K)

I see there's still autism-tier, quality content on Jow Forums
include me in the screecap please

>I cannot make it an open-sourced extension
>I don't want to help you. I just want to get rich from this app that I made for myself. Pay me goy

hack the planet bro, have a (You)

>t. ancap

Attached: 1545345583816.png (960x960, 331K)

build a front end app, then run the backend on a lamda or dedicated ML box on AWS

>with openCV
Can OpenCV implement neural networks?
I mean I used to work in an office with a neural network guy, and all I ever heard him talking about was TensorFlow, PyTorch and Caffe.

(I used to work on 3D geodata, and I've used OpenCV for texture mapping from aerial imagery onto 3D building models.)

You’d have a much larger impact if you open sourced it

Haha, life just got way more boring. Gratz.

>not being able to see behind OP's thread.
>thinks he needs this app to help him filter out the 100's of girls that will match him.
>genius
>most of you are

reminds me of that web app that identified racial percentages from a photo, but was pulled down due to sjw reasons. today, op was not a faggot.

I'd rather scream at women on the street until I get arrested than use Tinder.

yes, very interesting.
tell us a bit more about how it works

Bump for your hard work

The lower the better obviously. Being Tinder the numbers will be high but as a general rule anything over 20 means that the women has lost the ability to pair bond and isn't worth having a long term relationship with.

Let me make this simpler on you faggot, if you have six pack abs your match rate shoots up by 1000%

If you have nice selfies and pictures that look like an Instagram male chad, your match rate shoots up by 500%

If you have both, your match rate is unlimited and you match almost every girl you swipe on, with permanent 99+ status

If you don't have any of those, and you upload your "selfies" and you only have a few selfies of you looking retarded like almost 90% of males on Tinder, you will not get a single fucking match or you will get the odd trash match

It's all about optimizing for visuals.

Do it.
An autothot detector would be good too for casual fucks.

if you are willing to settle for unreliable data such as jawlines and the such it can be done.

And you're an idiot.
What's the purpose of being intelligent if you don't use it in any way?

Thats nice, but it will probably be forbidden, because (((tinder))) wants goblins and std-thots.

Read it again. OP said he uses Tinder as a source for pictures because he can get so many, but the project is not restricted to Tinder.

> working towards a PHd
>creates app for cash flow, names it E-ugenics

Your are either trolling or criminally autistic. Putting aside the fact that you should probably check what is your university's stance on PHd candidate entrepreneurship (you might be required to get their consent and / or pay a sizeable fee), you will basically destroy your own future forever.
Your best bet is to repackage it as some generic service with a "wink-wink, nudge-nudge" toward the possible applications, and charge access to other entrepreneurs who are less vulnerable.
This way you can broaden your audience and not take any heat.
..While writing this, I actually started liking you. I'm also a CS major, currently doing my own thing. If you'll do what I suggested, it will be a good service to other developers, by making this sort of image processing accessible and easy to use. Do you have a demonstration to show us?

I don’t have much money but I want to donate to you, smart Dane user.

Nigger, what?

so where's the git

And here comes Shekelberg. Better not steal his idea, kike

I realize it's not going to be perfect but it would be pretty amusing if there was *some* positive correlation between the results and reality. Roastie butthurt would be phenomenal if it takes off.

show your software in action query traits get results

Why are jews always so dumb?
how do you think this tech works answer this and ill change my opinion
new computer science fag

that would be quite amusing actually

Just use Typed Arrays and pass the matrix calculations to WebGL, then it won't be a problem running it in browser noob

This, give us descriptions of images in this thread

Attached: gib-gf.jpg (780x1600, 282K)

This is not how it works.
If software can be compromised by having the idea stolen, it's shit.
Implementation and exposure is what gives it value, and that cannot be stolen unless he shares the code.

It would be less humiliating to be literally try and rape a girl in broad daylight and get arrested in front of my friends and family, and have that be broadcast on national news, than to create a Tinder account.

I don't care about your opinion :^)
Learn to make conversation.

Uh ... anything using a neural network that does NOT run on GPU kernels is factor 1000 to 2000 slower. So instead of waiting maybe 2 seconds for a single result, you'll be waiting 40 to 80 minutes.

so you are a cs major and you dont know how it works?
typical from a jew

It would be great if the same technology could be applied across all major normie dating apps/sites like okcupid.
Or perhaps Dane user could use his skills to create a new rival dating app that allows users to set racial preferences and filter results accordingly.

The problem with the specialized "White Only" political-themed dating sites like white date is they are so niche, hardly anybody knows about them (or are put off by the political angle), so they're usually 95% dudes.

I think the thing to do instead of that niche route is to have a completely normie-friendly platform that people of all backgrounds could use, but would allow individual users to filter results to any preference they like (including racial and political preferences). That way it would have the popularity and user-base of the large apps/sites, but would allow the user to essentially use it like the specialized race-based niche apps.

People like us would essentially be piggy-backing on the normie masses to further our own objectives.

It's popularity with those normies could be maximised if it didn't include the profiteering traps that most of the major apps now have, or the bots, scammers, etc...

Attached: 1515672977074.jpg (1024x615, 184K)

WebGL runs on your GPU

But OP already said how it works. He uses a GAN and scaled/rotated images from Tinder with known traits to let it train itself.

You are going to fail super hard at life if you continue with this.
OP is an autist, but he has merit and managed to not fuck up all the way to a PHd.
You're just an autist, a rude one, and I'm not wasting my time on you.

i think opencv does with work neural networks
would be shit if it didnt

docs.opencv.org/2.4/modules/ml/doc/neural_networks.html

I was using OpenCV for Harr cascades a decade ago but it was slow as shit

Yes, all right. But I'm not so sure about the potency of smartphone GPUs. Like we tried at the university to get a detection software run on a tablet PC and we failed. Run time was 1.3 seconds on a PC with a GeForce 1080 and around 40 minutes on the tablet.

Unless you have a very flat/thin neural network, you're going to run into hardware issues pretty soon.

>thinking you have to pay a fee or get consent to make your own business and code just because you're also studying for a phd
hahaha the university doesnt own him and his research unless he is specifically working for them