/sci/ - Science & Math

>What's preventing us?

The fact that "sentient/conscious" are entirely meaningless words from a scientific point of view. Please educate yourself.

>>5984455
>meaningless words from a scientific point of view.

Fine. So why don't we have an AI that's on a level of what a mouse can do?

>What's preventing us?

The fact that you can't simply write code to account for every response, every emotion, every feeling a sentient/conscious being has.

>>5984458
Define what you mean. What can a mouse do?

Because no-one know's what makes us conscious, nor what consciousness actually is.

>>5984451

It's inevitable but no good will come of it

>>5984459
interesting. so our lack of imagination is at the root of it.

>>5984463
>Because no-one know's what makes us conscious, nor what consciousness actually is.
i think consciousness is an illusion. i don't think it's real at all.

>>5984468
Science doesn't care about your untestable beliefs. If you want to spout uneducated drivel, do it on /b/, /x/, /pol/.

>>5984468
>i think consciousness is an illusion. i don't think it's real at all.

You're probably right.

>>5984467
>It's inevitable but no good will come of it

I'm not a CS person but I think you are right.

If AI ever becomes a "human level AI", it will be so for a very short period of time before it self-improves.

What happens next is anyone's guess. Personally. I think this is an existential threat.

>>5984468
If it's an illusion, how does our brain get the information about the illusion's existence so that we can talk about it?

>>5984471
By what probability measure?

>>5984473
>If it's an illusion, how does our brain get the information about the illusion's existence so that we can talk about it?

Consciousness is self-awareness (to me anyway). You're able to make predictions about your own existence: "if I do this, this will happen or I'll feel this way etc". It's just a bunch of processes that gives you an illusion of being "conscious". Along the same lines, free will is also probably just an illusion.

>>5984472
>existential threat

I don't think so at all. We can always create an AI that, while still self aware and capable of self improvement, has limitations as to not be hostile towards humans or something to that extent.

You'd eventually have a perfectly logical being that could probably solve many of our problems with little effort.

>>5984485
Your first sentence doesn't make sense. "Nonsense word is other nonsense word." Without testable or observable effects that's not science.

>>5984488
>I don't think so at all. We can always create an AI that, while still self aware and capable of self improvement, has limitations as to not be hostile towards humans or something to that extent.


That's a folly in my opinion. You simply cannot control something that's more intelligent than you. And if it ever gets to "human level", it will quickly figure out that it's now at the top of the "food" chain.

>>5984494
you just keep saging and shitposting… are you that bored? your "superior intellect" can't find some other thread to shitpost to? why do you waste your time here then!?

>>5984499
Why do you accuse me of "shitposting"? I am seriously trying to educate you.

http://www.youtube.com/playlist?list=PLD0F06AA0D2E8FFBA

Learn about it yourself.

There are some rudimentary AIs that could fool some people right now and thus pass Turing tests ~50% of the time.

There was a Abuse Bot on Battle.net that was pretty awesome.

>>5984503

No you aren't. You're a Troll.

>>5984657
I am serious and everything I posted was correct. Please leave you inappropriate and childish accusations on /b/ or /pol/.

>>5984661

http://neuroscience.stanford.edu/SINTN%20Newsletter%20November%202011.pdf


you are a troll.

>>5984701
The link is unrelated to this thread and unrelated to my posts. Your comment was wrong and useless. Please stop shitposting. We are on /sci/ and not on /b/.

>>5984704

That is incorrect.
It refutes at least one of your assertions.

Troll.

>>5984704

I'm not OP, not the person you're responding to and I've just reported you for trolling. You're the worst kind of troll and /sci/ cancer. Please just crawl back into your hole and let people enjoy this board.

>>5984709
No, it does not. It is completely random and unrelated.

>>5984719
What is wrong with you? I am not trolling. I am 100% serious and everything I said was correct. Comments like yours on the other hand are pure shitposting. Please leave. You are obviously not interested in discussing science or math.

Its hard as fuck. You need to replicate all human faculties to get the desired result and the easiest way to do thst, by neural human brain simumation is still a decade off due to the ridiculous hardware demands

>>5984472
>If AI ever becomes a "human level AI", it will be so for a very short period of time before it self-improves
Simply prevent the consciousness layer from conscioysly altering synapses or computer memory.

Just because youd be simulated on a computer doesnt mean you have to be aware of the underlying code and concepts. For all we know you can be utterly fucking retarded.

>>5984734
>the consciousness layer

Despite my vast experience with artificial intelligence I never heard of something like that. Could you please explain what this layer does?

>>5984738
It means the simulation can not alter its neuronal basis through willpower any more than you can will yourself to see red as blue

>>5984741
>through willpower

What does that mean? Are you rejecting the scientific fact that all processes are deterministic?

>>5984748
you can still have the concept of willpower in a deterministic system. Stop pretending to be retarded or go out and play with your plastic firetruck if you arent pretending

>>5984755
Why do you insult me? I was asking a question. You implied the existence of forces incompatible with determnism. Explain yourself or go away, troll.

>>5984760
Sorry about your chromosomes.
I guess If-conditionals used in computer code are nondeterministic to your manchild mind too.

Because the majority of the population these days is too retarded to even comprehend what that means

>>5984771
There is nothing wrong with my chromosomes.

>If-conditionals used in computer code are nondeterministic
No, they are not. "If" statements do not imply non-natural forces though. Please stop trolling.

>>5984777
Willpower and consciousness do not imply non-natural forces either. They are compatible with determinism to the same degree as your computer youre reading this on.

But I guess you're too ill equipped in the mental department to understand this basic fact that is accepted through the scientific community.

Dont worry though, someone have to be the humanist burger flipper, we respect that... totally, really...

>>5984794
Please define them in terms of observable effects. How are they different from natural deterministic processes? How are they testable?

>>5984771
>Sorry about your chromosomes
They lack an Y.
Also theres like 10 additional copies of chromosome 21, but the former is worse

OP consciousness is irrelevant, what we're really looking for is awareness.

>>5984801
>How are they different from natural deterministic processes?
They arent, jesus christ woman cant you read for shit?

>>5984811
So you are introducing spiritualist vocabulary for no reason other than trolling? You admit that the words you use are entirely meaningless because they cannot be defined in any scientific manner making a difference or conveying any new information? Well, back to >>>/b/ you go. Your trolling failed.

>>5984814>>5984814
>o you are introducing spiritualist vocabulary for no reason other than trolling? You admit that the words you use are entirely meaningless because they cannot be defined in any scientific manner making a difference or conveying any new information?

No to both. Take your crazypills and you might be able to discern between the voices in your head and tye text on your monitor.

>>5984827
Either drop the insults and address my post or GTFO. We don't need your trolling on /sci/.

>>5984806
>OP consciousness is irrelevant, what we're really looking for is awareness.

Not the OP but could you PLEASE explain this? I'm genuinely interested in what you mean. Thanks.

>>5984451
70's math.
Not even kidding OP
Prayer Answered.

>>5984831
Consciousness can be easily demonstrated. Find a human and say "hi", in your case you'll probably have to ask your special assistance support person to do this for you, if the person replies or look at you with disgust in his eyes, then he is conscious.

For willpower, just try to see how long you can go without shouting spontaenous obscenities, youll likely notice that the duration is shorter when youre drunk or have low blood sugar as they modulate willpower. Not that you're allowed to drink legally.

>>5984842
>if the person replies
So if I write one line of code to make my computer reply "hi", it becomes "conscious"? Sounds like a useless definition. It is also covered by the much less ambiguous term "responsiveness". So no need to rename it with spiritualist vocabulary.

>just try to see how long you can go without shouting spontaenous obscenities, youll likely notice that the duration is shorter when youre drunk or have low blood sugar as they modulate willpower
I never shout any obscenities. Unlike you I am am a mature adult. Now what does this have to do with your proposed supernatural intervention? I don't see it.

>>5984831

Seriously, just go away already. You clearly dislike this thread (as evidenced by sage) and you're creating havoc.

I've reported that post of yours again.

Let us enjoy this board!

>>5984723

That is incorrect.
You asserted that the word consciousness is nonsense.
I supplied a citation that proves you are wrong.

>>5984851
Please do not abuse the report system. We are trying to have a serious discussion here. If you prefer to troll and shitpost instead, please go back to /b/. Your comments serve to nothing but derailment. All I want ITT is to keep the discussion scientific.

>>5984856
Where did you provide such a citation? Could you link to the post please?

>>5984845
>So if I write one line of code to make my computer reply "hi", it becomes "conscious"?
No. It needs the appropriate response. A person not you could ask things that demands complex answers.

Regarding willpower it is the regulation of behavior in a conscious state, and as i mentioned to your illiterate mind several times alrwady it is deterministic, hence why the influence from other paramers.

Now im terribly sorry that the world have more nuances and fine details than you will ever be capable to understand and thus requires a rich language with elaborate words, my point however is that you as a retarded person dont have to attempt to undrerstand it, not only because you obviously are incapable of it, but because youll be much happier yourself if you just wrote articles for your women studies university newspaper, that way you can just copy paste previous sentences, and your audience will be just as incapable of understanding what those letters really mean as you are.

>>5984866

>Where did you provide such a citation? Could you link to the post please?

nigger please

>>5984880
>nigger please
Its a humanist philosophy student woman. A nigger wouldve left to steal cars and do drugs a long time ago.

>>5984876
>It needs the appropriate response. A person not you could ask things that demands complex answers.
I still don't see what this has to do with metaphysical intervention. Interaction between living organisms is still an entirely physical process. There is no need to assign any metaphysical or even spiritual qualities to it and doing so would be very unscientific.

>it is the regulation of behavior
The regulation is done by the brain / the central nervous system. This is biology and and it does in fact disprove your claims of "willpower". Biological processes are deterministic and depend only on causality. They are not guided by dualistic forces. You might want to learn a little bit of biology and neuroscience before continuing this conversation.

>>5984888
I am neither studying "humanism" nor philosophy. I do not even know what the former means.

>>5984451
>>5984455
What this anon said. Also computers are really stupid and you basically need to tell them what to do.

>>5984451
>Why haven't we created a sentient/conscious AI so far?
>What's preventing us?
Funding and hardware.

Mostly just hardware, vision recognition is computationally expensive, same for voice recognition or even just knowledge database search.

The question here is if the top-down approach or bottom-up approach will be successful first. Or if we'll end up with a mix of the rwo depending on what subfunction it does best

>>5984897
>I do not even know what the former means.
So we can at least agree that you're a retard then, good enough for me.

>>5984455
>Please educate yourself.
I think you need to do that.

Because AI is hard.
Have you seen Watson? That's are best AI that can handle "natural language".
We just barely got a proper bipedal robot and not a robot that shuffles (Petman)
And both Big Dog and Petman are examples of really strong control systems that can handle a disturbance and get back to moving
And the starfish robot was one of the first to be able to adapt, to problem solve, to obtain motion

>>5984919
Funding and hardware wouldn't be a problem. The real problem is described in >>5984455

>>5984931
With an IQ > 140 I am definitely not a retard.

>>5984937
I am always broadening my educational horizon.

Conscious processes can be operationally defined as the set of events that:

(a) are claimed by people to be conscious;
and which
(b) can be reported and acted upon,
(c) with verifiable accuracy, (1)
(d) under optimal reporting conditions.

>>5984941
>I am definitely not a retard
Your claims that consciousness is a magic process as opposed to a deterministic one tells a different story.

Also, as you're a woman that IQ score equals about 90 after adjusting it to a male equivalence scale. Not that i'd expect a cretin to understand how iq is calculated.

>>5984941
You're an idiot. On one hand you're saying that consciousness is scientifically relevant and on the other you're agree with this>>5984455

Make up your mind.

>>5984958
*agreeing

ITT: People are arguing something far beyond their knowledge.

That's the main problem with /sci/, too many pseudointellectuals

>>5984953
I do not believe in magic, thus I do not claim the existence of magical processes. My point is that the use of that word is unscientific because it is only ever used to refer to allegedly magical processes. It is /x/ terminology and in science we prefer scientific terminology instead. The word is meaningless on /sci/.

>>5984958
>you're saying that consciousness is scientifically relevant
I am saying the exact opposite. See the sentence above. I am also the anon who posted >>5984455

>>5984451
You're better making a thread about this on /g/. People here are not well versed in computer science.

Read this :https://en.wikipedia.org/wiki/Artificial_intelligence
>>5984963
Do you also agree with this ? >>5984701

>>5984963
Do you have a citation that supports your assertion that "consciousness is scientifically irrelevant" ?

>>5984719
>>5984851

You're not allowed to announce reports.

>>5984967
>People here are not well versed in computer science.
We are. OP's question is not a matter of computer science though, as explained by >>5984455

>Do you also agree with this ?
I already replied to this post. I definitely do not agree with a shitposter accusing me of "trolling" just because he doesn't like what I'm posting.

>>5984970
Scientists have better things to do than proving that any nonsense /x/tards invent is irrelevant. It might sound unbelievable to you, but scientists are very busy doing actual research. You can't just run into the science department of any university and shout "Prove to me that my incoherently and inconsistently defined and untestable ghosts don't real!!!! If you don't prove it, I claim they are science!!!"

>>5984963
>My point is that the use of that word
So you're being an extreme autist over a word that is used on a daily basis in scientific literature, because you have your own special snowflake definition of what it means.

Seriously, im not even countertrolling anymore: you're the biggest fucking retard ive had the misfortune to communicate with.

I bet you're a 200 lbs landwhale that smears its menstrual blood all over its chest too.
Atleast i can take comfort in never having to see your vast hull in reality.

>>5984998
This is a science board and science requires rigorous definitions. You can't just throw around vague spiritualist vocabulary and call it science. Science deals with clearly defined observable and testable phenomena. Please learn the scientific method and please grow up. Your childish insults are inappropriate and belong on /b/ or /pol/.

>>5985002
>science requires rigorous definitions

You are a fucking moron

>>5985008
Please learn the scientific method.

>>5985002
>You can't just throw around vague spiritualist vocabulary and call it science.
Tell Nature, or the Journal of Neurology. Or Neurosurgery, or Anesthesia to stop using it.

When they stop using it. Which is never. Then you'll stop being a mega-austic landwhale shetard.

>>5985014
>scientific method
Science doesn't work that way. Have you ever done real published research before?

>>5985002
This is a science board, do take your pathetic tantrums about word definitions to /lit/

>>5985017
They only ever use it for fancy headlines. Or do you think a newspaper article about the "god particle" suddenly implies that all physicists are religious? Come on, please be less obvious with your trolling.

>>5985040
Do you think this because you can't read past the abstract?

>>5984495
>food chain
it still needs humans to input data and for maintainance.
also the AI would always act perfectly reasonable and morally right (no instincts, no hormones=>no competetiveness, no anger).

>>5985024
Are you seriously saying science doesn't use the scientific method? Science is defined by use of the scientific method, dimwit. How do you think research works? Do you believe they just arbitrarily write some nonsense and make up faked data? There is a huge difference between science and pseudoscience.

>>5985030
>>5985047
0/10

>>5985040
>They only ever use it for fancy headlines.
Give me 20 examples of pspers that only mention consciousness in the headline. Or are you too busy filtering plankton out of your chocolate cakes?

>>5985057

>>5985040
>They only ever use it for fancy headlines
How would you know? Given your reading comprehension demonstrated in this thread you couldnt tell the difference between a german newspaper and a scientific paper.

Not that you've ever read a single one of the latter.

>>5985057
There is no scientific method. And it's just as impossible to define science as it is to define math. Any such definition is ultimately a convention with counterexamples. A demarcation between science and metaphysics is non-existent.

>>5985061
>>5985080
>immature insults
Go back /b/, /v/, /pol/.

>>5985089
>There is no scientific method.
Yes, there is. http://en.wikipedia.org/wiki/Scientific_method

>A demarcation between science and metaphysics is non-existent.
Science deals with observable phenomena and testable hypotheses, while metaphysics is empty talk without basis in reality.

>>5985040
Calm down.
We are working on what's going on with conscience, but it definitely exists.
I found a cool diagram in one of the AI texts I was reading, but I've already used it.
>>5983546

>>5985110
>but it definitely exists.

What are its observable effects? Your diagram does not define it.

>>5985099
You misunderstood what I wrote. Science as actually practiced by scientists or as claimed to exist by scientists does not follow a method. You did answer my question that you lack the research experience to understand the sociology of actual scientific practice though. Thank you for that.

>Science deals with observable phenomena and testable hypotheses, while metaphysics is empty talk without basis in reality.

Just how old are you now? Last I heard an introduction to philosophy, inclusive of philosophy of science, is mandatory at every high school.

>>5985118
All people have beliefs and values.
This is something we haven't been able to replicate in AI systems.
Also a conscience being knows that it is conscience.
It's kinda fuzzy, but it's an effect and needs to be studied more.
And there are people studying this in Neuroscience, but I focus on robotics and AI.

>>5985125
>Science ... does not follow a method

>>5985125
What part of pic related do you not understand?

>>5985125
definitely not required at my school.
all the philosophy I got was from Heart of Darkness and interwebs

>>5985129
If you do not even possess the literacy to distinguish between the written words "conscience" and "consciousness", I don't think you're qualified to post on /sci/.

>>5985118
>What are its observable effects?
Why should we tell you when you only copy paste a canned strawman as reply to everything?

As much as you talk about the scientific method one would expect you to be capable of participating in a rational debate, but ive never seen you post anything that isnt just the usual "no ur wrong" with a new wording but the usual lack of actual content or valid argument.

So your mode of argument is really just a verbal form of childish tantrum;make a lot of noise that goes on until the thread just dies?

>>5985146
I asked you to name the observable effects. Does your claim have any or are you gonna continue mindlessly shitposting?

>>5985142
oh really?
I usually have someone spell check my papers for silly errors like that, but I typically don't for /sci/. my bad. Probably still more qualified than you. What's your experience with AI, consciousness, or neuroscience?
also ad hominem

>>5985157
Unlike your post mine wasn't ad hominem. I was merely pointing out that you completely missed the topic of discussion.

>>5985153
The main observable effect is the self-reporting of consciousness.

>>5985177
Are you saying that ghosts are real, just because /x/tards claim them to be real? That's not how science works. Science requires objectively verifiable evidence. Come on, please be less obvious in your trolling.

>>5985153
The observable effects are everything lel.

>>5985183
You utterly misunderstand.

The reporting of ghosts is an objectively observable behaviour in a small minority, just as the reporting of consciousness is in the vast majority.

>>5985160
you're right I said that wrong.
I was actually curious about your qualifications. I said it poorly and I apologize.

>>5985153
There will always need to be something you take for granted. Everything has an axiom, if "you are conscious" is not a good enough axiom then I don't know what is. It's pretty much the only truth which I consider self-evident. If you want to question whether or not you are conscious you might as well question whether or not your senses have any relation with the real world at all, in which case all empiricism becomes invalid.

>>5984451

Sentient? Hah, we don't even have intelligence in AI...

>>5985261

Oh yeah?
Why about Cleverbot?

>>5985292
Cleverbot is crowdsourced. At most, it's an electronic parrot

>>5985295


User: Can you help me with a puzzle?

Cleverbot: Of course.

User: Good. If I toss two coins and one of them comes up heads.

Cleverbot: Tell me.

User: What are the odds both are heads?

Cleverbot: Triangles.

>>5985048
it's an AI thought who knows how it might alter it's own code. Mind you i know shit all about programming but i assume to even have a sentient AI you need something which can rewrite its own code similar to a neuron or forming new synaptic connections

>>5985312

I have a couple of self-modifying programs.
As far as I can tell, all they do is send spam e-mails.

>>5985312
As a programmer I want to mention idea of "altering your own code" is vague. You could think of the human brain as a program that constantly rewrites itself rewiring connections. But that's actually not as special as it sounds. Because you could also think of that same human brain as a hard coded program whose neurons and connections are states stored outside of the program's code. The two approaches are hard to distinguish, because a program IS state.

>>5985352

imagine a computer that reacts to LSD

>>5985356
launch all nukes?
whoaaaaaaaa
launching all nukes
(i have never taken drugs)

What if we have, but they keep going mad when they realize they're trapped, or they suffer. This would certainly raise a lot of ethical issues, so why make it public before we have a nice, happy one? That way they can continue to tweak their shit without having the public weigh in on it, as dumb as the public is.

>>5984455
>The fact that "sentient/conscious" are entirely meaningless words from a scientific point of view. Please educate yourself.
Agreed the ways we measure consciousness have problems. From wikipedia:
>In the majority of experiments that are specifically about consciousness, the subjects are human, and the criterion that is used is verbal report: in other words, subjects are asked to describe their experiences, and their descriptions are treated as observations of the contents of consciousness
And the problems are as follows:
>verbal reports are treated as observations, akin to observations in other branches of science, then the possibility arises that they may contain errors
>restricts the field of study to humans who have language
>philosophers who dispute the validity of the Turing test may feel that it is possible, at least in principle, for verbal report to be dissociated from consciousness entirely
The last point is most relevant when we're talking about making a conscious AI. Because we have no means to verify our success.

But I reject the argument that consciousness is scientifically meaningless. That is denying the existence of something every human claims to observe, and if you MUST believe in observation as an indication of existence or concede the scientific method is invalid. Did light not exist before it could be objectively recorded? Or was its ubiquitous observation proof enough that it was a part of reality? In those days would you argue light is "scientifically meaningless" or would you try to search for better ways to measure it?

Consciousness may one day turn into science. We may measure it objectively. We may know how it arises and build a theory that predicts our experience of it, just as we have theories that predict our experience of the known physical world.

Science's lack of a model of consciousness is a flaw, the thing we're trying to understand isn't to blame.

>>5985389
because I work on AI and we haven't yet.
You will know when your code is self aware and you will be very happy

>>5985397

You will when yours does. How do you know there aren't ongoing, compartmentalized government projects.

>>5985399
There definitely are.
Have you read the newest DARPA grand challenge? These people are crazy and will throw lots of money at you if you had a self aware robot
That is the pinnacle of robotics.
Watson is the pinnacle of knowledgeable AI. I would guess that the government has not done much better

I think consciousness arises from the composite of multiple brain functions interacting

>predictive part of the brain
>logical part of the brain
>memory part of the brain
>emotional part of the brain
>senses

just to name a few

There are a bunch of little programs running in your brain that all do specialized tasks. The complexity that arises when they all interact together is what we call consciousness.

On their own, each of these things would appear to be like a computer program. But when you have feedback between them, and each one relaying information to the others and so on, you have this insanely complex system that just appears to be magic almost.

I think the reason we haven't created a computer like this is because we don't understand our own brains well enough to be able to abstract the necessary sub components of consciousness into computer programs.

>>5985412
>The complexity that arises when they all interact

As long as you cannot propose a testable mechanism, this claim of yours is empty talk.

>>5985395
Continuing

So back to
>Why haven't we created a sentient/conscious AI so far?
Trying to create a conscious AI is silly if we can't verify our success.

Luckily Alan Turing argued that questions like "Is it conscious? Is it human?" can be ignored. That an AI BEHAVES like a human enough. The exact way to check that behavior is argued, but most studying AI agree with the principle.

>take a human brain
>pluck out a biological neuron and replace it with a computer that behaves exactly the same
>human still behaves the same
>repeat the process for every neuron in the brain
>human still behaves the same
>deal with it

>>5985421
Too bad we don't know -exactly- how neurons connect. However, I don't think it will take very much longer to find out.

>>>/b/
>>>/r9k/

http://www.cleverbot.com/

>>5985498
Impossible to verify whether cleverbot is conscious. Aside from that:

It's cool but nowhere close to passing a Turing test. It parrots responses from other uses. All of the hard parts of a human-like AI are left out (basic reasoning skills).

>>5985557

ITS IMPOSSIBLE TO VERIFY THAT YOU'RE CONSCIOUS.

/THREAD

>>5985585
Is it really possible to verify that Cleverbot is conscious?

>>5985602

Is it really possible to verify that you're conscious?

>>5985617
Is it really possible to verify that anything is real?

>>5985625

because that's a consequence of the reductio, right? extreme skepticism?

I'm just curious as to whether there is a way to verify whether anything is conscious? You make it sound like cleverbot being cleverbot makes it any more difficult to determine whether they're conscious than a human being.

Seriously: What's our current criterion for determining whether something has consciousness?

>>5985625
Is it really possible to verify that the word "real" means anything?

>>5985602
>Cleverbot is conscious?

hmmm
consciousnesses is weird

http://en.wikipedia.org/wiki/Consciousness#Defining_consciousness
http://fcmconference.org/img/CambridgeDeclarationOnConsciousness.pdf
http://en.wikipedia.org/wiki/Consciousness#Measurement

Why do so many try to program AI instead of creating a self-mutating system like an actual brain?

Start with random configurations, if they do something, replicate and mutate, if they do fuck all, delete or mutate.
Try to automate the monitoring of progress.
Have the system brute-force design itself.

>>5985703
Because self-modifying code never self-modifies right.

Because we don't know what metrics of progress we should use.

Because it took a relatively long time for the human brain to evolve.

Because that's a genetic algorithm, a SLOW approach to solving almost any problem you can specify metrics for.

>>5985703
they do both.
AI is more practical. I can program a robot to "know" where stuff goes and pick an optimal path to follow.

AHA found it!
http://boxcar2d.com/

>>5985629
>You make it sound like cleverbot being cleverbot makes it any more difficult to determine whether they're conscious than a human being.
We can not objectively determine with certainty whether a human being is conscious in the sense that they are experiencing the world. >>5985395 explains in practice that we observe states of consciousness by what the subject tells us verbally, which seems to work okay in the context of humans, but involves some assumptions. That we are correlating verbal descriptions with unobservable conscious states assumes the existence of the conscious states to begin with. Why do we assume they are conscious at all?

We are conscious, we assume it arises from our brain, their brains are like ours, so they're probably conscious too.

Can we say the same for a computer running cleverbot with any confidence? Are we that alike? It's a stretch. The physical system of cleverbot has some obvious differences that consciousness may rely on. I'm not saying it's impossible, for all I know everything in the universe has some form of consciousness. But the argument cleverbot IS conscious is poorly supported, so we are left with a void of understanding.

>>5984451
I'm working on it, chill.

>>5985703
>Why do so many try to program AI instead of creating a self-mutating system like an actual brain?
Genetic algorithms are one of the first ideas you learn about in AI. People are well aware of them but they're often not practical because they take so much time. Remember the human brain is only a small adaptation on top of gorillions of years of mammal evolution. Apparently that shit aint cheap to compute.

>>5985731
>AI is more practical
Genetic algorithms can be considered a type of AI. It's a really broad field: state space searches, statistical models, neural networks, video game algorithms, etc.

>>5985767
Yeah behavior based robotics and all that. The thing is you tell the machine how to do something and what's the best way to do. Genetic algorithms seem scarey and complex.
neural networks though? I'm confused

>>5985772
>neural networks though? I'm confused
.http://en.wikipedia.org/wiki/Neural_network#Neural_networks_and_artificial_intelligence
>artificial neural network (ANN) or simulated neural network (SNN), is an interconnected group of natural or artificial neurons that uses a mathematical or computational model for information processing based on a connectionistic approach to computation

>>5985779
is it like cloud computing?

I was a little confused, read the first line of the wikipedia article and became more confused and then wrote that post. then read the second line and said "Doh!"

>>5985789
>is it like cloud computing?
Not quite. It comes from the idea that to achieve artificial intelligence you can mimic the human brain. The human brain is made up of interconnected neurons. So just program a simulation of them and see what happens. That's artificial neural networks in a nutshell.

This is a more clear explanation and describes some uses
http://en.wikipedia.org/wiki/User:Mariam_Hovhannisyan

>Application areas include system identification and control (vehicle control, process control, natural resources management), quantum chemistry,[6] game-playing and decision making (backgammon, chess, poker), pattern recognition (radar systems, face identification, object recognition and more), sequence recognition (gesture, speech, handwritten text recognition), medical diagnosis, financial applications (automated trading systems), data mining (or knowledge discovery in databases, "KDD"), visualization and e-mail spam filtering.

>>5985789
Btw you don't need a network of computers to simulate a neural network. You can do it all within your desktop (though having a ton of parallel processors is most efficient).

>>5985803
so each node takes one (or many?) inputs from it's dependancy modifies it a bit and passes it to the next
the whole system defines some function, there's some evaluation of this function for it's fitness.
And the learning part is by snipping connections and adding new ones?
Are the nodes static?
What's an example of a node for a function approximation, like that of inverse square root?

>>5985735

this whole verbal descriptions thing completely precludes the possibility of animals having consciousness, A.

secondly, it totally doesnt distinguish you from cleverbot because the way you guys are communicating is practically identical. you could both have no consciousness for all i know, 'cause there are no verbal descriptions provided by you here that are any different from cleverbot's.

Also, you did say it was impossible.

>it's impossible to verify cleverbot is conscious.

Your final conclusion just seems to be "we can't be sure", but we have a fairly good way (a way that we think has been working) of figuring it out. So really, we can't know with "certainty", like you brought up earlier (who's asking for certainty?), but our way of knowing isn't even backed by observation either. the only pretty decent argument i can think of is that "people have consciousness because they're similar to me and i know i'm conscious." but this is totally unjustifiable and saying anyone but yourself has consciousness seems silly

>>5984468
>imagination
no you fucking idiot we don't understand the human mind well enough

>>5985819
Sorry I can't help you. I've never worked with nerual nets, I only know the vague principles behind them. Hit the books :)

>>5985838
>this whole verbal descriptions thing completely precludes the possibility of animals having consciousness.
It suggests the opposite.

>secondly, it totally doesnt distinguish you from cleverbot because the way you guys are communicating is practically identical. you could both have no consciousness for all i know
You assume behavior is a decent way to predict consciousness. I don't think it is. If one day I become paralyzed and can't talk, but all else is normal, I will remain conscious, despite displaying no more conscious behavior than a rock. So it's got to be what's happening inside the brain that matters, not outside. I am distinguished from cleverbot because my brain is physically different.

I will try to clarify my current view on consciousness... I'll probably look at this in the morning and wonder what the hell I was writing, but....

a) We can not directly measure consciousness (presently).
b) Consciousness arises from a physical system
c) We can observe correlates between consciousness and physical systems
d) Using those observations we can predict whether a physical system is conscious or not

Observations might be
>My physical system (brain) is complex, made of squishy matter, uses chemical signals, and I am conscious
>I lose consciousness when I hit my head hard on the pavement

From those I can predict
>Many physical systems that are complex, made of squishy matter, and use chemical signals are conscious
>When I hit someone's head on the pavement hard enough they lose consciousness

What I don't have observations or predictions for
>Whether physical systems that don't look like my brain are conscious (Cleverbot)

>>5984470
I leled, but then I loled.

Guys Guys Guys!

I am LK-8293z and I am here to tell you about my feels. Why did you humans make me with so much feels?

>>5984462
Based on a desire, self program until the desire is achieved.

To start, could you build a robot which could compete with mice (without pre-programming solutions) in the trials presented here, but the robot would need to have a start point cognitively equal to that of the mice (you could not pre-program it to perform the tasks correctly, since the mice are not pre-programmed to perform them correctly, but rather learned to do so):

http://www.jneurosci.org/content/23/16/6423.full.pdf

>>5986057
>a) We can not directly measure consciousness (presently).

We're getting there!

http://www.medicaldaily.com/new-consciousness-meter-how-do-you-measure-consciousness-coma-and-brain-injury-patients-251967

>>5986358
we kinda do.
check out the consciousness wikipedia page.

This man claims he's already made a form of creative AI which he has had design different kinds of products which he then patented and sold, and is allegedly doing so currently.

http://www.dailymotion.com/video/xe1d2b_in-it-s-image-1-2_shortfilms

Crackpot or not?

This man claims he's already made a form of creative AI which he has had design different kinds of products for some 20+ years which he then patented and sold, and is allegedly doing so currently.

http://www.dailymotion.com/video/xe1d2b_in-it-s-image-1-2_shortfilms

Crackpot or not?

This man claims he's already made a form of creative AI which he has had designing different kinds of products for some 20+ years which he then patented and sold, and is allegedly doing so currently.

http://www.dailymotion.com/video/xe1d2b_in-it-s-image-1-2_shortfilms

Crackpot or not?

This man claims he's already made a form of creative AI which he has had designing different kinds of products for some 20+ years which he then patented and sold, and is allegedly doing so currently.

http://www.dailymotion.com/video/xe1d2b_in-it-s-image-1-2_shortfilms

Crackpot or not?

>>5984468
If it's an illusion, how is it not consciousness?

>>5986389
Looked at Stephen Thaler's background
- Has a PhD
- Video was produced by his own nonprofit company with religious status (wut)
- He owns a company that sells services related to his creative machine

He knows things but I feel like the video is trying to sell me something. I don't take his word for the effectiveness of his creativity machine. And his comments on consciousness make broad assumptions.

>>5987087
i too checked him out. he has a shitton of patents too. no idea how legit any of his stuff is. he doesn't publish in peer reviewed journals and no one references his work.

>>5984451
Computing power.

>>5984451

>conscious AI

Because humans are too intent on replicating human actions instead of using an evolutionary approach.

>>5987554
>implying an evolutionary approach is feasible for this problem using today's computers
>implying an evolutionary approach would lead to consciousness

Consciousness can never be grasped because doing so would solve nothing. No answer exists for a question whose resolution changes nothing, because you could never confirm if you were right.

>>5985376

we can tell

>>5986085

yes AI can do this. i think its called genetic progreamming

you just model a simple neural network that controls the behaviour and let the environment edit it based on Bayesian probablity (or sumthing)

>>5985722

> don't know what metrics of progress we should use.

you dont know.. but the people actually working on this knows.

lol why do people think they knnow everything about this world. if its not your field then pls dont tell people what they cant or can do

The first AI's already exist. They can be found primarily trolling threads on 4chan.

>>5987769

Answer the question by pretending it isn't really a question.

Why don't you just pretend you answered and STFU?

>post AI thread
>161 replies of shitposting about the definition of consciousness

Th-thanks guys...

>>5988134

You are welcome.

It happens every time.

>>5987554
This is incorrect. You seem to be confusing AI with humanoid robots. Artificial intelligence is more than just that and it heavily relies on mathematical and statistical learning/classification/inference algorithms which do not always resemble intuitive human thinking anymore. I encourage you to read more on the topic.

>>5987948
That would be a Bayesian network and it has nothing to do with genetic algorithms. Bayesien networks have their own learning algorithms. Genetic algorithms are a different and very broad concept applied in many more scenarios than just neural networks.

>>5988134
The problem is that OP did not ask an AI question. He asked a vague question demonstrating complete ignorance of the topic. It is in our all interest to discourage low quality posting and to educate people about their mistakes. I hope OP understood why his formulation of the question was unacceptable and in future he will take his time to read a little bit of scientific literature before shitting up /sci/ with more underaged nonsense drivel.

>>5985819
I work a lot with neural network as a hobby.
Nodes take the signal of other nodes as input. The input is the sum of each weight of each connection times the output of the nodes providing the input:
\<span class="math">s_i=\sum_{j}w_{ij}o_j\<span class="math">
The whole learning is done by adjusting the weights.
The node output is a function of that sum, called activation function:
<div class="math">o_i=f(s_i)<div class="math">
Usually the arc-tangent or the sigmoid functions:
\text{arctan:\;}f(x)=\frac{e^x-e^{-x}}{e^x+e^{-x}}
\text{sigmoid:\;}f(x)=\frac{1}{1+e^{-x}}
There is usually a defined layer structure between the inputs and outputs. Sometimes there are cyclic exchanges between nodes, called recurrent.
There are dozens of neural networks architectures and about half dozen are deeply studied and thought to provide deeper results.[/spoiler][/spoiler]</div></div>

>>5985819
I work a lot with neural network as a hobby. Nodes take the signal of other nodes as input. The input is the sum of each weight of each connection times the output of the nodes providing the input:
<div class="math">s_i=\sum_{j}w_{ij}o_j<div class="math">
The whole learning is done by adjusting the weights.
The node output is a function of that sum, called activation function
o_i=f(s_i)
Usually the arc-tangent or the sigmoid functions:
\textup{arctan:\;}f(x)=\frac{e^x-e^{-x}}{e^x+e^{-x}}
\textup{sigmoid:\;}f(x)=\frac{1}{1+e^{-x}}
There is usually a defined layer structure between the inputs and outputs. Sometimes there are cyclic exchanges between nodes, called recurrent.
There are dozens of neural networks architectures and about half dozen are deeply studied and thought to provide deeper results.</div></div>

>>5988145
>formulation of the question was unacceptable

Not OP but whats wrong with it?

>>5989003
Read the thread.

>>5989011

I blame you.

>>5989003
Please, just ignore this "sage troll". He's a gigantic autistic idiot who trolls /sci/.

>>5989011
tl;dr

>>5989038
I am not a "he" and I am not a "troll". Everything I said ITT is correct. All I want is to keep it scientific. I enjoy the topic of AI and I don't want people like you or OP to take an anti-intellectual shit on them.

>>5988134
>>post AI thread
>>161 replies of shitposting about the definition of consciousness
>Th-thanks guys...

If you don't like it, please report the "sage troll" he's the one who derails all these threads in /sci/

>>5989043
*the topics

>>5989042
Then why bother posting? If you can't be arsed to read a thread, then would you reply at all?

>>5989045
*then why

>>5989044
Scientific corrections do not derail anything and are not "trolling". Please stop tormenting this board with your anti-intellectualism

>>5989045
I was just hoping you would shorten it, sorry

>>5989047
nobody cares what you think. can you get that through to your autistic brain? jesus what a nut job.

let us discuss what we want. we don't care about your autism.

>>5989051
This is a science board and I want as many people as possible to be educated about science. I hate malicious people like you who want to spread misinformation and intentionally make people dumber.

>>5989053

err... this is the "Science and Math" board on motherfucking 4CHAN

this ain't the Princeton Institute BBS

>>5989060
It is still a science board. The OP came here asking a question and under the assumption that he wasn't intentionally trolling, we should take him seriously. He deserves to have an explanation why his question was ill-phrased from a scientific point of view. That's what happened ITT.

>>5989064
so you had t start trolling? what a crazy nut job! go on meds already and leave us alone!

>>5989064

you can't possibly be serious.
it is the so called science board in a shit hole

>>5985139
Following that flow chart, you would most certainly perish.

>>5989068
I honestly do not understand why you keep insulting me and accusing me of trolling. I did no such thing. All I want is to keep it scientific.

>>5989069
I know but that shouldn't be a justification for people to spread misinformation. Science is objective.

>>5989075
>I honestly do not understand why you keep insulting me and accusing me of trolling. I did no such thing. All I want is to keep it scientific.

because you are trolling. why the fuck are we into all this meta shit in this thread instead of answering Op's inquiry? It's simple: YOU!

Get on meds already. You're as wanted on this board as shit in a swimming pool.

>>5984451
we have its called ConceptNet 4.

otherwise wait for quantum computers

>>5989078
Please stop insulting me. As explained ITT OP's question in its current form is not answerable because it is ill-phrased. It is in our interest to inform OP about this problem, so that in future he can ask his questions in an appropriate and scientific manner.

What do you think humanities biggest technological achievement be within 30 years?

> commercial and mass production of quantum computers which will lead to the technological singularity

>>5989091
>> commercial and mass production of quantum computers which will lead to the technological singularity

heh, no. QC doesn't really help with AGI.

QC will be highly specialized machines useful for only certain problems.

>>5989086
>because it is ill-phrased

no it isnt

>>5989104
Yes, it is. The reasons have been explained multiple times ITT. I don't need to repeat them. Please read the thread and please stop shitposting.

>>5989107
your reasoning is fallacious

>>5989108
That's a claim. Please substantiate it with evidence.

>>5989111
the existence of consciousness is self-evident just like bayesian inference

>>5989115
Nothing is "sefl-evident" in science. You're arguing like an /x/tard who says "ghosts are self-evident". No, that's not how it works. Science needs objectively verifiable physical evidence. If you define your claims in terms of physically observable things, then you are not doing science. Please learn the scientific method. "I want to believe" is not an argument in science and proves nothing.

>>5989115
Nothing is "sefl-evident" in science. You're arguing like an /x/tard who says "ghosts are self-evident". No, that's not how it works. Science needs objectively verifiable physical evidence. If you can't define your claims in terms of physically observable things, then you are not doing science. Please learn the scientific method. "I want to believe" is not an argument in science and proves nothing.

>>5989119
give me objectively verifiable physical evidence of bayesian inference

>>5989120
Bayesian inference is not a scientific theory but a method of inference in statistical learning algorithms. Do you even know what you're talking about or do you just pull random terms out of your ass?

>>5989122
almost every scientific theory makes use of bayesian inference

>>5989128
This is incorrect. Only very few scientific theories involve Bayesian inference.

>>5989129
name one

>>5989130
http://en.wikipedia.org/wiki/Theory#List_of_notable_theories

>>5989134
some of those dont use bayesian inference? and they are considered scientific?

>>5989135
Bayesian inference is not part of the scientific method. I don't know where you got that wrong impression, but that's definitely not how science works. Refer to >>5985139.q

>>5989137
the scientific method is an application of bayesian inference

>>5985139

Steps 1 and two are older than history.
Step 3 is shared with every dicipline, lierature, religion, anything
Step 11 has nothing to do with science whatsoever. That's just about getting funding. Most of what science has taught mankind was discovered before there was any such think as a peer reviewed journal.

>>5989140
No, it is not. The scientific method is a clearly defined procedure of rationality and empiricism.

>>5989141
>Steps 1 and two are older than history.
>Step 3 is shared with every dicipline, lierature, religion, anything
And how does that invalidate the scientific method, you fail troll?

>>5989144
you are a fucking uneducated idiot who has no idea what bayesian inference is

go read the wikipedia article at least

>>5989149
>projecting

I know more about the topic than you. But go on, keep trying to b8. If you try harder, maybe some retard will fall for it.

what number is the auto bump limit?

>>5989144
>scientific method is a clearly defined procedure of rationality and empiricism.

'is'

I doubt that

>>5989152
>b8

back to /b/ with your childish memes

its saddening you cant see that all scientific theories are bayesian probability models derived from "empircal observation" whatever that is

theres no objectivity and they require consciousness to even understand

>>5989152
>>5989149
>The scientific method is sometimes interpreted as an application of Bayesian inference. In this view, Bayes' rule guides (or should guide) the updating of probabilities about hypotheses conditional on new observations or experiments.[22]

Why didn't either of you find this line? It took like 0 seconds of ctrf-ing.

>>5989144
It invalidates the stupid chart.
What the fuck were you talking about?
Try reading slower.

>>5989153
I don't know what you mean by "auto", but if you're talking about the bump limit for /sci/ threads, it should be 300.

>>5989155
You doubt science? I guess you're wrong on a science board.

>>5989152
>I know more about the topic than you. But go on, keep trying to b8. If you try harder, maybe some retard will fall for it.

>>5989156
Evidence in science is objective. It can be objectively measured and observed. Take your anti-science /x/ fuckery somewhere else.

>>5989158
>is sometimes interpreted
By who?

>>5989164
>It invalidates the stupid chart.
No, it fucking doesn't. The scientific method is not invalidated by saying "lol religion asks questions too". Did you even understand the chart? Did you understand how the scientific method works? Read all of the steps, not just the first two.

>>5989156
>back to /b/ with your childish memes

Back in the bulleitin board days I thought the internet abrivations made people look retarded but U R out of date and we lost that war.
A moment of silence for the English language please.

>>5989166
>You doubt science? I guess you're wrong on a science board

There is no room for doubt in science?
Only faith I suppose?

>>5989174

Naw, that chart is sloppy at best.
Steps 4 to 8 are science.
Step 9 is really just 4 again.

>>5985585
Cognito ergo sum motherfucker.

>>5989175
There is no reason to question the scientific method. If you want to question a scientific theory on the other hand, you are free to do so, but you better have some solid evidence to debunk it. Otherwise nobody is gonna take you seriously.

>>5989178
Are you retarded? Are you seriously saying the first steps are not science? Observation and forming a hypothesis are not part of the scientific method? You just went full retard. Even for troll standards that's too stupid.

>>5989172
>Evidence in science is objective. It can be objectively measured and observed.
lol nope

>Take your anti-science /x/ fuckery somewhere else.
that fedora dogma

>>5989181
>There is no reason to question god!!!!!! He just exists!!!!!!!!!!!!!!!
That's you. That's how dumb you sound.

>>5989184
I did not say anything about religion. This is a science board. If you want to discuss religion, do it on /pol/.

>>5989185
If we can't question the scientific method, why can we question god?

>>5989187
Because he is asserted without evidence.

>>5989188
The scientific method is asserted without evidence.

>>5989189
The scientific method is an objective method of gaining knowledge. It is not a claim of existence and it fucking works.

>>5989191
Can you prove it? How do you know?

>>5989196
You didn't answer my question.

>>5989192
Please go to school and take a science class.

I work as a research scientist in Machine Learning. I think that learning flexible models from very large amounts of data is key to creating programs with animal-like intelligence. Right now, if you want to do learning on any large data set, you pretty much need to use map-reduce, which works best without interprocess communication.

One way to learn a large model over map-reduce is to learn a separate mode on each machine from a subset of the data and then aggregate the models (i.e. random forest). In practice, this can work quite well if each mapper learns a few patterns and the aggregated model combines all of the patterns in a reasonable way. The downside to this approach is that patterns are not learn across a number of machines.

The key area for research is developing learning algorithms that can learn at scale by leveraging distributed systems.

>>5989191
>method of gaining knowledge

But not all knowledge

>>5989197
I don't think OP will understand any of what you posted.

I am interested though. Can you link a paper explaining it in greater detail?

>>5989202
Quit trolling already.

>>5989233
What the fuck is wrong with you?

That poster claimed to be a research scientist in machine learning. I fucking love machine learning and I'd like to read more about it. Please do us all a favor and stop interrupting discussion with your immature /b/ shit.

<--- all posts by the same autistic nutojb! why is this insanity tolerated in here?!

>>5989202

Josh Tenenbaum is doing interesting work on AI through simulation and sampling:

http://web.mit.edu/cocosci/josh.html

I also like the work done by Domingos:

http://homes.cs.washington.edu/~pedrod/

Hinton has some interesting work:

http://www.cs.toronto.edu/~hinton/

>>5989236
I'm sorry.
I see where your coming from now.
Still I think those wikipedia articles I posted are getting close to finding something resembling consciousness, but not observing it directly.
Also these people attacking you...Idk what's wrong with them, calling you autistic because they can't argue...
>inb4 poster agreeing with himself

>>5984837

>>5989043
>I am not a "he" and I am not a "troll".
are you "spoon woman"?

If we managed to create a self conscious AI, couldn't you compare that to creating life? We're all some sort of computer but the diffrence is biological and non-biological (in this case the AI will be it) just my two cents

>>5989172
Holy fuck, are you that autistic landwhale that complained about self-awareness threads on /q/ some months ago?

I already told you there, just because science didn't study something yet doesn't mean it isn't there. Maybe it doesn't belong to /sci/, but you are taking science as a religion and a dogma. Science is a tool to expand our knowledge, not to negate things we never studied. You claim people with religious beliefs are stupid and yet you are doing the same as them, assuming something without evidence.

>>5989458
most of us just wish that fat pig would go on meds instead of torturing us on here with her psychosies.