The J-AI returns

The J-AI returns
Shalom. I think it's about time Zig Forums made a funny app to redpill some Zs and trigger some kikes. Current plan is to do transfer learning on a facial recognition network then optimize it with tf lite, and cook it into an app. It's basically like make-app but instead of removing makeup it gives you a jew-score. If Zig Forums has any data scientists, you'll have to excuse my ignorance since ML/programming is a hobby to me and this project is meant as a fun little project. I will need data for this, since we're aiming to detect tiny and large, ethnic facial characteristics and this requires some precise and well sorted data.

When enough data is gathered I will take advantage of (((googles))) goodwill, and use their free colab platform for model training. Oy vey.

Any recommendations, criticism of oversight on my part or good ideas is of course welcome. But right now the key phrase is: DATA, DATA and MORE DATA

Attached: py.jpeg (646x272, 14.62K)

Other urls found in this thread:

archive.vn/XNuuN
archive.vn/KPZCB
gitgud.io/J-AI/
researchgate.net/publication/318138434_Prediction_of_Human_Ethnicity_from_Facial_Images_Using_Neural_Networks
kairos.com/)
gitgud.io/J-AI
twitter.com/NSFWRedditGif

If this thread doesn't die I will post some updates later when I get some sleep and more data. Or I'll just make a new one.

I will go through archive.fo to find the old threads
archive.vn/XNuuN
archive.vn/KPZCB

Not bad, that should trigger them. Here's another idea along the same line of thinking.

Create a news type aggregator app that pulls the author's name and then searches whether or not they are a kike. You can even build a huge kike database for future purposes. I mean they want a gun registration database, I don't think they would oppose a kike registration database right? ;7 Most jews will identify as jewish publicly for the virtue signalling/never forget the 6 million, so I'm sure the scraping isn't terribly difficult.

Shit fucking thread. I thought the actual J-Dar page was back.
OP a nigger.

see it is another Jewdar project that is coming back.

oh and if possible, maybe make a graph to create metrics of people they work with to get a jew collaborator score if they themselves are not jewish

There is no project here. OP is an ideas guy who is looking for people with actual employable skills to do the dirty work for him.

sage negated

Attached: jew post.gif (480x270, 1.05M)

I only asked for good data sources, where the fuck did I state that any work was going to be unloaded? I simply stated that more knowledgeable people exist on Zig Forums, and that they are welcome to call me a faggot and whatnot. This shit has just been created, don't expect it to get done in a matter of minutes, especially data gathering and verification are slow ass processes, which is why I made this thread.

Attached: screenshot-drive.google.com-2019.01.03-12-25-13.png (1174x233, 12.1K)

J-Dar was fucking great.
Ironically it was originally made to promote how great Jews are in entertainment, yet was used almost solely to know whether to avoid a film or not.

Bumping this for Zig Forums

Oy vey, OP here. I'm currently creating some scripts for sorting through the data gathered in earlier threads, it's about 21k images in total. When I get the scripts up and running I will split the data to enable interested anons to donate 10 mins to classify some images. QUALITY is imperative since we are trying to detect pretty subtle facial features. I also tried compiling an android app with a model that detects faces, so far it seems nice and functional.

Attached: cad983acb210180b162c1faa544d524bb36a880d35c0dea1507c7af07d418209.png (442x558, 49.94K)

Please deliver in gitgud.io if you can

Shalom

Just bump

Coming up, I'm running some distro upgrades on my VM and laptop and I'll upload some packs with the script, images and pip requirement files for the script

Bumping for the repo, extra points if you can start a thread in >>>/poltech/

bump

Attached: crypto.jpg (750x940, 100.38K)

Is that board active? Oh well, I will need about 40 people doing 500 images or 20 people doing 1000 images each.
Repo is at gitgud.io/J-AI/
Nothing has been uploaded yet but i will get the 40 zip files and accompanying scripts up.

It does not work, I tried it last year.

Contrary to popular believe here on pol the facial features of jews and gentiles are rarely distinguishable, jews and whites are mixed too much. The currently available models are not sophisticated enough to pickup the tiniest differences, nor is there a non biased dataset available of sufficient size. Creating one of sufficient quality is a rather impossible task. Tracing back through all pol threads will net you with sets of something about 10k and 5k iirc. Its not only dirty, biased and unbalanced, but its also not nearly enough. You will have better accuracy looking at a side view of the nose and last name.

TLDR your personal/manual judgement will be more accurate than a trained model would

Then how is it most people can tell the difference in one glance? If a person can do it, a machine can do it 100 times easier.

From my current training results it seems like the data is the number 1 issue here. When the model loads a very well sorted batch it reaches 80+ accuracy, whereas a bad batch yields around 50-60%. The 21K images I have aren't sufficient to train a fresh model, but transfer learning on a net trained for facial recognition seems very possible. Modern models are incredibly versatile.

Because we take in much more information as input than just a face, such as the model will. Moreover, facial feature extraction models are not advanced enough to pick up these rather negligible differences. Attached images are from the jewish dataset, some might be obvious some might be not at all, in a next post i'll attach some from the gentile set. Apart from very obvious cases, most people fall in the gray area where it is incredibly hard to distinguish between jew and gentile based on this alone.

It is practically impossible as of now, though theoretically its possible.

Attached: jews.5721_0.jpg (155x154 11.04 KB, 9.69K)

Just warning you now that you're wasting your time, though its a fun learning project

Attached: gentile.379_0.jpg (108x107 3.96 KB, 4.51K)

Computers will never replace humans, but it will enhance them. This could be a good crutch.

Which model did you use? And how is your data obtained?

Idk user the first 3 in that image are obvious. Especially the 2nd and 3rd.
You are probably looking at images of post-plastic surgery merchants if you think they are indistinguishable. I highly suspect the 4th one had his face altered.

I already get a feeling someone might be jewish just looking at their face even if they have no curly hair and have always been right, if A machine can predict homos on facial features with something like 80% accuracy, it can find a jew too.

It's 90% my dude

transfer learning with vgg16 (even tried with inception, resnet and some others for lulz), tried it with multiple subsets of data collected from prevoius pol threads, manual collection, twitter/vk crawlers.

Had little to do with facial features, it was based on selfies without any real preprocessing so angle, lighting, maybe colors, accessories, presentation etc would all impact the outcome

researchgate.net/publication/318138434_Prediction_of_Human_Ethnicity_from_Facial_Images_Using_Neural_Networks

98% for a CNN. I know this is major ethnicities, but getting even 75-80% is more than enough for the purpose of a funny app used by Zig Forums and teenagers having fun. I currently use images that often contain most of the hair and ears at 224x224px and like you said yourself, I think the problem is in the data, since Zig Forums is pretty good at detecting jews and current NNs achieve human level performance on most facial recognition tasks.

read my previous posts again
you will never even reach 55% unless you use some terribly biased dataset consiting of nothing but caricatures at which point the entire thing is redundant anyways

J-dar was a lot of fun. If only there was a way to pull its index and mirror it.


Bump. As you said it would be worth it just for the sake of letting other people use it.


If you're so adamant, how about you release your version and let OP use it as an example? If it's as you said, a waste of time, then you could help OP save some time so he can begin his attempt at it. Maybe he will find something you missed. Help him by helping him save some time.

You yourself mentioned caricatures, but even admit an experienced Zig Forums user can detect a kike. The less pronounced features however are still detectable. That enough is a way of helping a machine detect kikes.

MOST jews aren't difficult to identify using physical features, voice/nasal tone, and behavior, which is the most telling. It would be nice to have an app to help those who need it until they can spot jews without it. I hate seeing people post Lauren Southern or some other jew when I can they are jews. Why can't most people see it? That's what needs to be addressed.

Why not an app called 'Your Ugly Self' that turns you into a super nasty looking kike? It would be fun for zoomers and a warning at the same time.

You're trying to classify those of jewish descend based solely on their facial features. Of course that means you can also make a distinct difference between those who aren't jewish and those who are. A problem encountered during the classification task of jews is that their features cluster a lot with other ethnicities. Most Jewish that are of Ashkenazi descend have mixed features, some predominantly European / Caucasian. It was hardest to classify those of Caucasian descent since there was the most diversity of facial features in those groups. You could split it up into eastern and western european features and you could still get a very high accuracy of 90%+ (that coupled with other racial groups).

The reason why caricatures are mentioned is that the classifier itself is only very reliable in regards to Jewish caricatures, or the Seraphim Jewish group. It's not really good at all to identify mixed race Jews or Ashkenazi Jews which are the majority of those who have Jewish blood in them. Due to the large overlap in clusters that are formed to classify, the accuracy when it comes to the majority of Jews is quite low.

The reason why caricatures are bad is because you will give the machine learning algorithm a bias to only detect caricatures. We don't want an algorithm that can only classify those who look like caricatures. The reason for that is you put in bias (caricatures) that are an obvious cluster of their own and then say 'look it works on other caricaturist looking jews'. But it wouldn't detect a Jew of Ashkenazi descend who has some European features. So it's rather useless because Jewish people are so mixed in particular with Caucasian people that they are very difficult to detect.


You can indeed as a human see some difference because you notice a faint mix of Jewish / Arabic features and white features. The people who have those features are often Jewish. It's difficult for a machine to see the same thing in a mathematical vectorized way due to the high overlap with other races problem.

Attached: Cluster-2.svg.png (1200x803, 8.57K)

My trash, unsorted dataset gave me 64% validation accuracy on average. 50% is just random guessing, which would imply zero difference in general facial patterns. Do you believe that neural networks aren't at a level currently where they can detect two distinct ethnicities? Or do you believe the task of creating a high quality dataset is too difficult?

Couldn't this be solved through a more granular dataset? Wouldn't the main thing blocking the networks generalization be the large variance between the Jews. I think the issue arises when you try to generalize on both western mixed Jews and caricatures. I never anticipate it to be able to detect plastic surgery crypto jews, but most normalfags don't even know jews are an ethnicity, so even if we are only able to catch the not so white looking ones, its fine.

I like this idea, seems like it would be pretty funny.

Fuck off nigger. Do it yourself and come correct rather than making shit threads trying to con others into doing the work for you.

Why are you so angry? Is it because you have zero reading comprehension?

Israeli here, most people use instagram here. Let me know if I can help in any way.

Still nothing to add.

Stop the infighting, our enemies already has the technologies on hand to do it (kairos.com/) so my best advice is release all source code to GitGud like gitgud.io/J-AI did, and see what we can cook up.

I would recommend researching on "Facial Landmarks" and see what type of data we can get from that.


We are literally making it happen right now, so either let it happen or fuck off.


I need you to scrape all social medias for photos, including instagram, VK, facebook, Snapchat… see what type of data we can get. We need one set for Israelis and another set for Iceland, Norway, Finland and Visegrad Four Countries.

bump