Artificial Intelligence used to Expose Women's Bare Faces

Why the wall behind changed color?

wall had makeup duh

results from this ai must be taken with a grain of salt because the craters had a vested interest in making them look worse than they actually would.

>the craters had a vested interest in making them look worse than they actually would.
>the craters
But also, I'm not sure you can really argue that.
Fact is that makeup is used to improve your appearance, so by definition they will look worse without it.
What you should say is that it should be taken with a grain of salt because it is a software estimation and it has not been tested to any level of accuracy.

But the reactions are still absolutely hilarious.

>When did we start pretending non-sentient programs were "AI"?
It doesn't rely on explicitly programmed rules for doing what it does and learned it through training on datasets.
If you expect it to have an inner monologue, emotion behavioral routines, hopes, fears, and ambitions, and the ability to discuss the plot points of your favorite TV show with you I would argue you're retarded. Just because people have many different loosely related mental processes doesn't mean it makes sense for AI programs to be filled with fifty different unrelated behaviors. There is nothing wrong with writing AI programs that learn and implement a single sort of task. This AI program is meant to alter faces and it does a perfectly fine job at it. Nobody including the programmer who wrote that program knows all the details of how it's making these image transformations, the program itself is the only entity that contains the knowledge for how to do this. AI is the correct label for this situation.
If your complaint is that it's still ultimately learning based on deterministic cause and effect relationships, well our learning is ultimately based on deterministic cause and effect relationships too (the brain does not appear to be impacted by quantum decoherence - Max Tegmark wrote a nice paper a while back explaining why neuronal firing timescales don't work for that - despite the Penrose quantum mind bullshit).
We're going to see mass production of the already existing self-driving cars soon, and if someone told you twenty years ago they invented self-driving cars you would rightfully interpret those cars as AI. And the process that makes self-driving cars possible is the same basic process that makes this program possible, with both cases using training on datasets rather than explicit rules based programming.

Wow, Abby's still a very pretty lady without the facepaste.

>Tfw gf looks better without makeup
>first time I saw her in the morning I had to check for myself
> now slowly convincing her she doesn't need it
>She's using it less and less
> Feelsgoodman

It's just a bot to make anyone's face look shittier, regardless of how much makeup they have on their face in the first place. Put Abraham Lincoln in and it'd just make him look like shit too.
Does /r9k/ really see something this dumb as the "ultimate nirvana of goodness?" Jesus

>It's just a bot to make anyone's face look shittier
That's because not having makeup on makes women look shittier.
>Put Abraham Lincoln in and it'd just make him look like shit too.
How exactly would you implement a machine learning algorithm that both learns how to remove makeup from pictures where it was used but magically somehow leaves pictures without makeup used unchanged? If you trained it on a bunch of non-makeup pictures like Abraham Lincoln that would fuck up what it learned to do a good job altering makeup pictures. You can't have it both ways.

>it's a "brainlets vastly overevaluate what a deep NN can actually do" episode

The term has always been used for all intelligent agents. Just because you saw it in retarded, anthropomorphised sci-fi, doesn't mean it has any bearing.

Things like these just confirm again how stupid your average imageboard user is. Everyone just quickly jumps on the bandwagon in hopes of scoring some easy (You)'s.

This basically. This is just way too complex of an issue for ML to solve yet. Also you would need to train it on a gigantic set of really high resolution images to get somewhat accurate results.

>Fact is that makeup is used to improve your appearance, so by definition they will look worse without it.

It's used with the ITENTION of improving the appearence. Whether it does it or not depends on the person and the skill of the application of make up. It's also subjective.

>That's because not having makeup on makes women look shittier.

This is subjective. Personally unless the make up is masterfully applied (which most are not able to do) it looks unnatural and worse. Also if you have bad skin it will be visible through the make up anyways.

fucking retards. Its literally just a thin layer of makeup spread over bare face. You don't need quantum computing to draw over sloppy maskara line

you can feed it the same picture again and again and it will just turn the person into an abomination.
honestly looks like it just removes mascara, makes people pale and adds random imperfections

This is why AI has the potential to be so dangerous. Most people can't wrap their heads around how alien it is.

Since fucking forever?

>maskara line
>maskara
>line

Mascara is a quick-drying liquid applied to the eyelashes to darken them and provide a look of volume.

of course you know about makeup, cakeboy

Aw, someone step on your poor little wee wee? Go run back to mommy and daddy.

Eyelashes are considered make-up? Skin moisturizer? Soap? Seems like this thing just crunches the darks and defauls to a pallid skin regardless of the skin's actual natural color

you walk around with soap on your face?

>This basically. This is just way too complex of an issue for ML to solve yet. Also you would need to train it on a gigantic set of really high resolution images to get somewhat accurate results.
it just desaturates the face and adds a few freckles here and there.

you could probably get better results if you really wanted. notice how the background and eyes are also somewhat desaturated, although it still does a decent job of mostly adjusting the face.

Mascara and fake eyelashes are makeup

u sure that's a lady?

Who can tell anymore, really

youtu.be/IMMI8HWhqEc

I've seen Emma Watson irl. She certainly doesn't look like the on the left.

probably AI

...

maybe she had no makeup on?

Someone put this through the machine

Watch as the left is not recovered, but a new horrifying wrinkled hag visage is spawned instead

fuck, she's 5 times cuter without makeup

can someone explain to me the big deal out of this?
didn't we all knew this beforehand?

total qt to a whore.
is she a pornstar or something?

...

...

we have to go deeper

youre a virgin what would you know about makeup

Waist of computational power

Something something The Exorcist

Yes. Its just a shitty filter.

Pretty accurate if you compare first and last face.

not a virgin, just never around when they wake up :^)

IT'S LEVIOSÃÃÃÃAAHHNN

kek

fuck you AI no one ruins my waifu

nice meme kid-o

JUST

Thank you for your post. I'm studying philosophy and informatics right now and struggling to get my head around the emergent nature of cognition and reason. My profs are fucking me with their views. I myself am torn between reductionism and "emergence". Shit cannot be grasped just by using a priori knowledge. I feel like Wittgenstein is right and philosophy is dead. My intuition is telling me new impulses for understanding "being" will come from quantum theory and AI. If I understand myself correctly, freedom/consciousness can be "programmed" in a way that you don't know what the free individual is gonna do, as in: (the) God (of religion) creating man as free but being surprised/disappointed by Adam and Eve fucking up. Still, I looked up the things you mentioned and it was a great and helpful read. Hell, I'm so eager to learn out quantum theory and ML, but it'll have to wait after I'm done with my current degrees.

Yes, some women wear fake eyelashes with mascara or pay for eyelash extentsions.

SHE LOOKS BETTER WITHOUT MAKEUP
WHY DO WOMEN DO THIS

>learned it through training on datasets
Literally bayes filter. You're a retard.

on the right they look better

It's a filter, the only source of it being an "AI" is the app description. It makes everyone pale even if they are wearing light makeup and turns the blush they wear into rosacea. Because of the unnatural paleness, all the lines and folds in their faces are exaggerated and it completely removes eyelashes and a part of the eyebrows. It's not a true representation of what they look like without make up as shown here. The only people that think it's real are people who've never been around women for any meaningful amount of time or autistic people with a complete inability to read human faces

This. The thing holding back AI for years was lack of data driving scientists to get clever with how they cast problems/parametrized the features fed to algorithms to get the best bang for what they had. Unfortunately this sort of "feature engineering" imposed a lot of bias and rigidity that gimped a lot of more classical techniques. The real innovation in the past few years has been in allowing algorithms to perform inference/mapping operations while simultaneously constructing a feature space optimal for doing so in an entirely data-driven fashion independent from human meddling.

The result is models that are spectacularly effective but overwhelmingly black boxes. Philosophically, it reminds me a bit of anti-realism. Deep ANNs don't give a fuck about how you represent the world or eve if you claim to do so. Entirely outside the scope of human intervention, they have formed a highly nonlinear, statistical mapping of input to output that is entirely alien to us in its details, but nonetheless obviously quite effective in achieving the task it was trained to do. More cognitive scientists/philosophers should be seriously thinking through the implications of this.