The absolute state of "A.I."

The absolute state of "A.I."

Attached: lewd racy picture.jpg (2048x1589, 179K)

Other urls found in this thread:

youtube.com/watch?v=SA4YEAWVpbk
youtube.com/watch?v=piYnd_wYlT8
en.wikipedia.org/wiki/McCollough_effect
youtube.com/watch?v=DraA8WuI198
twitter.com/NSFWRedditGif

Nice cherry picked example

Im a human and this looks Adult and Racy from the thumbnail

Yes, attacks on neural networks are possible and you can create misleading examples even by hand. Imagine classification systems could (and will) do better. However, this is not a problem specific to AI but to all intelligence. Next time someone tricks you, think about it.

>Im a human

[doubt]

even "A.I." undestud the art
but not (You)

What's "spoof" for?
Post a screenshot of a solved captcha.

s/Imagine/Image/

youtube.com/watch?v=SA4YEAWVpbk

it's like a fuckton of random weights possess no intelligence at all. who woulda thunk it? (oh yeah the ai winters of the past...ssshhhh we need funding)

You are correct, _random_ weights possess no intelligence at all.

This. From distance it looks like abstract rendition of black chick being fucked.

Your brain is basically the same, m8

I thought the weights are no longer random after training.

but "massaging" them does...though information theory suggests random sequences have more information in them (see compression)

[citation needed] satan :^)

DNNs classifiers can get tricked by changing a single pixel in an image.
To trick a human requires much larger changes to image.

>[citation needed] satan :^)
The way individual neurons work is very well documented, and the brain is built up of neurons. You do the math.

>To trick a human requires much larger changes to image.
Yes, because our neural networks are much larger and better trained.

What does racy mean?

sexual in nature

Racially offensive.

Biological neurons and the brain is not equivalent in function to artificial neural nets used in deep learning. It's not simply a matter of scale.

And the brain has specialized structures for vision.

fuck it's getting late here, and I'm gonna miss this debate on the philosophy of ai and consciousness, in which 4chinz solves it once and for all lol

>Biological neurons and the brain is not equivalent in function to artificial neural nets used in deep learning. It's not simply a matter of scale.
They have some extra features like hormone receptors etc, and the structure is more three dimensional and dynamic, but yes it does work on the same exact principles.

It looks pretty racy, to be fair

That's sort of like saying a control system that uses transistors is basically the same as an AI running on transistors. There are worlds of complexity between those, complexity that matters. Sure, the non-random weights do things, but they lack one of the most important parts of our brain (relating to intelligence and making decisions/assessments): the ability to view and judge our own aptitude based on abstract outside stimuli, including our thinking itself. The weights can only do what they do, we can organize our weights and change them, or add new systems of weights.

There are tons of hardcore A.I. worshippers who seem to think if you have multiple different A.I. systems that are then aggregated and fed into another system and so and so forth for layers upon layers and have feedback systems between the layers then you eventually will get something similar to the human brain or even better.

They seem to forget the current AI systems only crudely biomimic only one facet of what's currently known about how a brain works, there are tons of known mechanisms that are not being adequately mimicked, and most likely, there are other mechanisms at play within the brain that we've yet to discover.

No you are what kids call NPC.

That's not a substantial difference. All the AI needs to improve itself is an internal subnetwork that is tasked with rearranging the weights of the rest and itself.

>racy
>humans keep feeding me this white vs black thing so I just call the colours like I see it

>All the AI needs

Yes?

well then get to it if that's all it needs.
bitches and money await

I never said it was easy.

competitive

It looks like a painting of a pair engaging in standing 69 ing

Non native English speaker here. What is racy?

This is a blue board, delet.

white female and a much bigger black male

its like a less autistic way of saying lewd

The absolute state of AI.
youtube.com/watch?v=piYnd_wYlT8

Attached: sddefault.jpg (640x480, 56K)

y'all ever think about how the brain is the only organ that named itself

you best stick to hip-hop, Tyrone

Looks like naruto and saske having sex together

>OP
>doesn't understand abstract art
>sees AI with clearly superior understanding of art
>OP shits on it thinking he is the smart one here

Talk about Dunning–Kruger effect

Attached: 7 - wAEWqft.jpg (270x374, 12K)

Neural networks are a dead end of AI development. People think of them as the be-all and end-all of machine learning because they are lazy thinkers and got fooled by seemingly promising results in the small scale.

The only way you can create an AI with a proper recognition of the world is if you generate a simulation of the real world in the "mind" of the machine. There needs to be 3D representation of a 3D object so the machine can actually "understand" what that object is rather than just (poorly) recognize shapes with a certain probability.
The machine needs to be able to run (many) simulated interactions with the simulated object to come to proper conclusions about the likely properties of the real life object. Then the machine needs to be able to confirm or deny the validity of those interactions that turned out to be successful in the simulation by repeating them as real life interactions. The real life object will probably behave differently than simulated. Those differences need to be recorded and translated into a revised simulation in the "mind" of the machine. Then the revised simulation runs additional tests with the simulated object, which then get tested again. And so on and so forth until a workable approximation of the real life object is created in the mind of the machine.

>Literally becoming so programmed by the kikes that your brain subconsciously perceives BLACKED when it doesn't exist
I bet you own an iPhone, huh

I got a boner from this picture

Jow Forums - neuroscience and psychology

t. Illuminati

Smarter than you, dork

Wrong. The brain is a nondeterministic turing machine due to QM. This allows the human brain to solve ridiculously hard problems in polynomial time.
Since probably P != NP, this means unless you can emulate a DNN on a working NTM (protip, we don't have any NTMs yet) then we can't emulate the human brain.

This.

I saw it as a black woman pegging a white woman in the thumbnail. The white woman is facing forward with her tits out, one arm reaching back behind her over the black woman's shoulder and the black woman is groping the white womans hips and has her head nuzzled into the crook of the white woman's neck.

Attached: 1534886597628.jpg (275x281, 31K)

Huh I'm seeing someone sitting cross-legged with his arms hiding his head

>en.wikipedia.org/wiki/McCollough_effect
The absolute state of humans

spotted the AI

imagine being this wrong

consider googling "do computer neural networks work like the brain" and come back to the thread when you're done

>[some] DNNs classifiers can [maybe] get tricked by changing a single pixel in an [edge case] image.

DNNs are a descriptor for a type of network typology. You can make a DNN which won't be tricked by a fucking single pixel change. Maybe when you when through your degree in India, the only DNN your peers would give you to stop you committing was one that was easily fooled

Both are Turing complete so theres a Brain to NN mapping

no way dude its definitely a black dude boning a white chick

I think we can all agree its both racy and adult themed, but its subject is up to interpretation

neural networks as a class is complete, but no individual nn has ever come close to mapping to a brain

using completeness to say 'the work on the exact same principles' makes no sense at all

>javascript and the human brain work on the exact same principles

>turning complete
>"yes it does work on the same exact principles."

these aren't the same thing. that's like trying to make a house only out of bricks, because bricks are used in making a house and you can 1:1 map a brick's function to a household function

Attached: 1.png (434x434, 81K)

>linux and the brain work on the exact same principles
>light globes and the brain work on the exact same principles
>cars and the brain work on the exact same principles

Everything is literally the same

Wtf I saw 2 anime girls kissing. One black haired and one white haired.
Has Jow Forums programmed me?

Attached: 516IfvA5N8L._SX341_BO1,204,203,200_.jpg (343x499, 27K)

A colored head amputee doing a reverse half-Nelson on a white girl while she performs oral sex on a golden retriever sure sounds adult and racy.

it looks like a blackman a latino women and a white women all having a threesome, the black man is eating out the white while the lationo gets the bbc

The absolute state of "humanity"

Attached: images(6).jpg (225x225, 4K)