Artificial Intelligence

[Compugasm]

Quote by: scofield Certainly the brain behaves in the same input/process/respond manner as any computer, but our uniqueness is in our potential to act outside of our programming.

Somewhat off topic, but now I finally understand why Neo doesn't die at the end of the Matrix. Thanks!

Quote by: Nemiroff The life force is energy. the food we eat, thats what keeps us alive. and we can simply replace that with electricity, or nuclear power, or whatever source we are using that far in the future.

I wonder how we would ever get on an airplane, if each person is more dangerous than a lithium battery?

Quote by: scofield The one thing that seperates us from being pre-determined automatons is the capacity to act, to effect without a cause. Some would call it free will.

A true artificial intelligence would be able to create art, and not merely duplicate brushstrokes, notes, or a scupture from instructions. One could argue, that fractal patterns are "art", however the computer doesn't appreciate it. Additionally, without the seed to create the pattern, it couldn't learn to paint the sky, or a tree, without the will to do so. Ok, that makes sense.

[Jubloz]

I haven't had time to delve too deep into this, but I read a short article in the new Discover Magazine (Jan '08 issue) about a project aimed at developing AI through evolving robots. Basically, the researchers had a group of robots, as well as "poison" and "food" to encourage a sort of natural selection. The "poison" drained the robots batteries, effectively killing them, whereas the "food" boosted their batteries. The programs of the robots who took the "poison" and were drained of batteries were scraped, whereas the survivors' programs were entered into other robots, some of which underwent a "random mutation" in their neural circuitry of 30 programmed "genes". By the 50th generation, the robots were actually communicating with each other; they were using their lights to indicate to other robots which batteries were poison and which were food. More amazingly--and this is the part that must've freaked the researchers out--one of the robot groups actually learned to lie. This particular robot group would tell other robots that poison was actually food, thus draining the batteries of the competition and allowing them exclusive access to the food. These robots had quite literally adapted to trick other robots for their own gain.I imagine that, if we see AI developed to any level that's substantial, it will be through a system like this that mimics evolution.

[iclaudius]

Wow, I am truly amazed to see so many dualists here today, and no one jumping to my defense. I thought this train of thought was killed off with post-modernism, but I guess not.

To the point, then: Probably the most common sentiment I'm seeing here adopts the belief that humans are inherently "better" than machines, that human intelligence is innately untouchable by artificial intelligence, and that in general people are -- dare I say? -- magically endowed with some special power of cognition that nothing we make can ever even hope to come near. Scofield, for example, responded, saying, "our uniqueness is in our potential to act outside of our programming" (in sentiment, anyway), that what separates us from our "pre-determined automatons is the capacity to act, to effect without a cause", and that we are not simply programmed to program ourselves. S/he seems to almost revere human intelligence.

Well, scofield, what do you think cognition is? The brain is basically a network of neurons, and those neurons are extremely predictable in the chemical responses they produce. How, exactly, do we act "outside" of our programming? In fact, how is that even possible? You do what your neurons are set up to have you do. Period. There is no alternative. We are clearly not set up to affect without a cause, and even if they were, how would that even be possible? Also, contrary to what you believe, when humans pick "random" numbers, their responses are much, much more likely to form patterns. When a computer picks a random number, it usually does so according to an algorithm picks a number statistically less likely to be picked than other options. Humans do not have the luxury of being able to calculate what should come next with complicated statistical analyses (not usually, at least), so they merely pick the number that seems least likely to them at that point in time. That is rarely, if ever, the most random number that could be chosen, and especially given the psychological effects of things like fatigue, their chances of generating something even remotely random are basically nil.

I understand that I cannot hope to dispel the notion of dualism completely here*, but what I hoped to address here is that the human brain is perhaps not as mystical and mighty as many of you believe. Scofield made the mistake of characterizing the complexity of human cognition as more complex than is physically (or logically) possible. And what s/he ends up demonstrating is that the basic tenants of AI and "actual" intelligence are the same; and that the biggest factors in our cognitive superiority are (1) level of complexion, and (2) self-realization. No, cognition is not made of anything particularly amazing. There are no magical levers being pulled in your head. There is no black magic going on. Your brain is a series of neurons that are programmed to fire in a very specific way. The difference between the standard computer program and you is self-realization.

That in mind, I'm going to directly address some of the things that have been said.

Quote by: scofield All programming is event driven. Set a computer running a program and it will run only according to the rules of that program. Sit a person down and they will generate thoughts without any external input or internal instructionset.

They will behave according to how their neurons fire. Human programming is still event driven, it's just on an incomprehensibly more complicated scale.

Quote by: nemiroff we are programmed to program ourselves? no no no, that doesn't make sense, perhaps you meant we are programmed to learn? programming is not learning. when you learn, you might not get the whole picture, you might understtand something wrong, different kids learn different things from the same lessons. when you are programmed, you just get the info hardwired into your brain. like brainwashing might be considered programming

And what is learning, nemiroff? Imagine this scenario: People are programmed to be curious. You stick your hand on the stove and quickly understand that it is very, very hot. You are now programmed to remember this whenever you think about sticking your hand on the stove, or actually, stoves at all. You have therefore essentially been programmed to generally avoid hot stoves. Brainwashing is a way of short-circuiting the learning process, but yes, it is still programming. You are essentially agreeing with me here.

Quote by: nemiroff I'M AGREEING WITH YOU, there is no difference between artificial intelligence and regular intelligence except the source. but if its made by man its called artifical. and the person with the replaced joint wouldn't call it artificial motion, he'd call it an artificial joint... the motion is perfectly natural unless he starts twirling it 360.

First, I'm saying "artificial" is irrelevant. If they're the same thing, how do you justify using 2 different terms? There's no point. Suppose everyone suddenly forgot what all the names of all the vital organs in the human body were, and you were in charge of naming them. Would you call the appendix by 2 names? No, of course not. Not without a good reason. Well, there's no good reason to differentiate them. Whether I make wood with nanomachines or grow it with a seed is irrelevant, so long as it's the same thing, and I can use it for the same things.

Quote by: compugasm A true artificial intelligence would be able to create art, and not merely duplicate brushstrokes, notes, or a scupture from instructions. One could argue, that fractal patterns are "art", however the computer doesn't appreciate it. Additionally, without the seed to create the pattern, it couldn't learn to paint the sky, or a tree, without the will to do so. Ok, that makes sense.

Why? An AI that could create and appreciate art would only need to be an AI that was programmed to be self-realizing and have an understanding of aesthetics, which (assuming you're talking about the human conception of the word) is dependent more on environment, values, culture, conflict, etc. A computer versed in these things would certainly be intelligent, but "true" intelligence does not entail this development. A computer could be as intelligent as a human being, for example, but not exposed to the same surroundings as us, and therefore not have the same definition of art as us.




* My boy Dan Dennett does a fine job of that in chapter 2.4, Why Dualism is Forlorn, in Consciousness Explained; the title is misleading, since consciousness is most assuredly not explained, but it is informative and more or less comprehensive in terms of creating a platform for solving the problem(s) of understanding consciousness.

[Kuldeep]

Quote by: Nemiroff we are programmed to program ourselves? no no no, that doesn't make sense, perhaps you meant we are programmed to learn? programming is not learning. when you learn, you might not get the whole picture, you might understtand something wrong, different kids learn different things from the same lessons. when you are programmed, you just get the info hardwired into your brain. like brainwashing might be considered programming

Learning is programming to the extent of data feeding only to the chips of MMM or to cells of human brain. But, the fuctioning as a result of that programmed learning varies from person to person wherein, genom mapping's role comes into picture, behavioural individuality for example. How and why genom mapping is different in billions of humen, in itself is impossible to understand !!! Therefore, tough task to duplicate human artificially seems to be also impossible.

i completely agree with 1 and 4, but how do you know they wont be able to display real emotion or develope a sense of individualisim? of course all the "AI" in games and basic robots and toys we have that display "emotion" are just prorgammed that way, but that doesn't mean they wont be able to in the future.

Displaying few emotions by countable sets of programmes and some mechanics does never mean one can develop innumerable number of emotions programmed to get inherntly existing within natural human being. As long as feeling of individuality is concerned it is even doubtful to ecactly feel, whether two humans feel their induividuality exactly same type. We get the idea by way of communication of thoughts, gestures and the like in unlimited number of manners and then we decide, alright all humans, animals, birds and insects have their unique identity.

I realy wonder if such type of dulicate emotions and individuality are develpoed in Robots in a situation, when we are not able to make JUST ONE naturally occurring ready made robot (HUMAN) deathless, eternally livable !!!

your right about the human body, the whole thing is a giant biological machine. ever seen scifi movies where they have one of those living ships? its not as tough as metal, but hey, our buildings will be able to fix themselves, produce our own heat. it isn't even that impossible, just breed certain types of cells and "construct" them into whatever shape you want... and as long as you feed it, you might even get an extra floor at the end of the year.

I do agree, few set of developments could be made here and there but, to replicate a full natural living being from a scratch is not possible. Even for making a Clon, scientists need living cell as life force ( not food as you have hinted in the following quote)

The life force is energy. the food we eat, thats what keeps us alive. and we can simply replace that with electricity, or nuclear power, or whatever source we are using that far in the future.

No, the life force, which I am refering is not energy but, the cause because of which cell is living. Food only helps to keep that cause continuing eternally living. That life force is not any of the physical matters or energies we are aware of. To my mind it is some meta-physical, we may term it Universal Consciousness.; the feeling of existence, which occurs both in living and no-living things as known to us commonly.!!!???? It is my wild guess. How can I prove meta-physical with physical tools ????

[Nemiroff]

you mean the original source of all life? seeing as the cause of all current life is the object that gave life to it. I'm guessing that random nonliving proteins combined with a wierd chemical fully atmosphere of the early earth, and massive amounts of insanely strong lightning storms triggered the proteins to replicate, (with certain mistakes) eventually creating mistakes that allowed it to move away from things that harm it, etc. etc, its a long story from there.

(This is a personal theory of one possible situation, i'm not gonna pretend that i'm talking science here, but its a possibility that has already been thrown around the scientific community.)

But the force that sustains life and that which keeps us alive is simply the energy we get from food. Meat eaters eat plant eaters, that eat plants, that absorb their energy from the sun.

First, I'm saying "artificial" is irrelevant

Its not irrelevant....its stating that people made it. All forms of intelligence to date are not artifical and to date we have not made any intelligence. as soon as we make some form of intelligence, it WILL be artifical.

That in no way implies anything about it's capacity, or ability to learn, create, feel, and even surpass us. It mearly states that we made it.

[Rainbow]

Quote by: DisQ Artificial Intelligence Is it possible?

I think it is possible to materialize the A.I. , but it would require :
- quantum computing, at least

[Kuldeep]

Quote by: Nemiroff you mean the original source of all life?

True, I am talking about the source which makes cell to function and grow to multiple. To me, that source is beyond all physical identities. Let me call is meta-physical beyond conception of our physical world. That source could be near to the origin of thought, Mind that too in meta-physical form viz, beyond neuronal impulse.

seeing as the cause of all current life is the object that gave life to it. I'm guessing that random nonliving proteins combined with a wierd chemical fully atmosphere of the early earth, and massive amounts of insanely strong lightning storms triggered the proteins to replicate, (with certain mistakes) eventually creating mistakes that allowed it to move away from things that harm it, etc. etc, its a long story from there.

I do not blame you. You can only think of physical concept but reality is meta-physical !!!

(This is a personal theory of one possible situation, i'm not gonna pretend that i'm talking science here, but its a possibility that has already been thrown around the scientific community.)

At least, I can not imagine how to view meta-physical with physical tools.

Power of thought or will power has been used by number of Indian origin saints. In physical terms we may give it a name as Miracle.

But the force that sustains life and that which keeps us alive is simply the energy we get from food. Meat eaters eat plant eaters, that eat plants, that absorb their energy from the sun.

Yes, that is true. It is the physical source (Sunlight), which sustains the meta-physical source, but not eternally....who can say why???? Aging then think why cells age ??

Its not irrelevant....its stating that people made it. All forms of intelligence to date are not artifical and to date we have not made any intelligence. as soon as we make some form of intelligence, it WILL be artifical.

That is correct !!!

That in no way implies anything about it's capacity, or ability to learn, create, feel, and even surpass us. It mearly states that we made it.

Science has infinite road to travel to reach to human mind, feeling and level of intelligence development.

[Charlatan]

I don't know much about AI, but do you mind if I build it up logically?

So you tell a computer it needs to survive, to stay functioning for as long as possible, and it needs to learn how to do that. If you take a computer you need to input basic features to get that going, like recognition of foods - batteries - like in one of the previous examples. The robot will need to regonise that food in his own way, but a deviation in the size of the battery would result in it not knowing what it is. Simply it needs to be told to keep functioning, and that to do this it needs the batteries. Can it move onto other places to look for batteries? Does it have this programmed into it? Which way will it go? With a complex program, it could be told that batteries are near to people, like a fly looks for food. So it will regard people as the providers of batteries, and could become a child looking for batteries. I don't think it will be able to look for batteries on it's own, so it will remain a child, looking to the people as providers of those, or something else that moves. So it will remain dependant on people to charge them for it, and to identify the place to go to get more batteries.

In making something learn, you need to tell it that anything that equals more existence is good, so that they have 'instinct'. In choosing what to store, it needs to store everything so that it remebers the bad and the good, that is a vital part of instinct, remebering things that don't work will lead to them avoiding those things in future, naturally. But how will it react to something new? It will either, based on previous experience, be curious or scared, as I guess those are two 'emotions' that will counter each other. So how do you make a computer curious? They will most likely seek out more battery power to consume, so will be more willing to forego their fear of eating poison and try the new thing out, seeing as how they needs that as part of their 'instinct'.

But how do you make it regonise new shapes? It has a video thing to recognise new shapes, and will be encouraged to investigate them because it needs power, so we are looking at artificial intelligence that is made to survive, as that is what intelligence is - the seeking of survival. If this thing seeks survival it has intelligence.

[Compugasm]

Quote by: Charlatan But how do you make it regonise new shapes?

If the intelligence doesn't interact with the physical world, it has no need for shapes. I was reading about the AI behind games like Halo, and the 'bots' learn and adapt to events, and assigning those events weighted values.

For example, running away from a hand grendade increases the chances of survival, as opposed to dying in the blast radius. The "learning" takes place at the next hand grenade thrown. Run away faster? Seek cover? What cover offers the best protection? Sheilds?, etc... as the senarios present themselves, the highest weighted values are attempted first. There are a good series of articles on AI Game Dev about building artificial intelligence.

[Rainbow]

Do you mean "Battlefield2" ?
I like that game :-)))

[Jack]

There's a rather disturbing story in the latest Wired about two young men who were working on AI.
Two AI Pioneers. Two Bizarre Suicides. What Really Happened?

[Compugasm]

Quote by: Isherwood Two Bizarre Suicides.

Uh, not these guys again. The idea of a giant databases to "mimic a conscious, thinking, feeling human being!" is, at best, a massive database of non-sequitor statements. IMO, I don't think it matters how many random facts it contains. This application isn't even close to AI. The only possible use I can see for this is to marketing or advertising agencies as a "chatbot".

Chatbot: If you are connected to an instant messenger, a chatbot starts a conversation with you. As the bot profiles your responses it will make a suggestion like "...oh, you might like this..." and then it links you to the homepage of whatever the topic of conversation is about. I haven't seen one of those since 2003. And I'm convinced Singh and McKinstry were behind that evilness.

[Apollo]

Quote by: Charlatan I don't know much about AI, but do you mind if I build it up logically? So you tell a computer it needs to survive, to stay functioning for as long as possible, and it needs to learn how to do that. If you take a computer you need to input basic features to get that going, like recognition of foods - batteries - like in one of the previous examples. The robot will need to regonise that food in his own way, but a deviation in the size of the battery would result in it not knowing what it is. Simply it needs to be told to keep functioning, and that to do this it needs the batteries. Can it move onto other places to look for batteries? Does it have this programmed into it? Which way will it go? With a complex program, it could be told that batteries are near to people, like a fly looks for food. So it will regard people as the providers of batteries, and could become a child looking for batteries. I don't think it will be able to look for batteries on it's own, so it will remain a child, looking to the people as providers of those, or something else that moves. So it will remain dependant on people to charge them for it, and to identify the place to go to get more batteries.

Several robots smiler to the one you're describing are already in existence, one of them is currently sitting on my desk. When the batteries reach a low level it will roll around in a random object avoidance pattern in search of sun to charge the batteries (solar powered) While its not of the caliber you're describing with human recognition and all that jazz, the principle is still the same, a self sustaining robot.

In making something learn, you need to tell it that anything that equals more existence is good, so that they have 'instinct'. In choosing what to store, it needs to store everything so that it remebers the bad and the good, that is a vital part of instinct, remebering things that don't work will lead to them avoiding those things in future, naturally. But how will it react to something new? It will either, based on previous experience, be curious or scared, as I guess those are two 'emotions' that will counter each other. So how do you make a computer curious? They will most likely seek out more battery power to consume, so will be more willing to forego their fear of eating poison and try the new thing out, seeing as how they needs that as part of their 'instinct'.

You're making leaps and bounds here, programming a robot with the ability to 'remember' where a food source is (we'll say sunlight for the sake of consistency) does not create a sentient, aware being. Every single thing that you're suggesting can be done with standard code, as for introducing a new variable to the environment you cannot assume a machine will show emotion. Chances are that unless a specific directive was programmed into the robot's memory to avoid battery draining devices, it will continue to function as normal (assuming it survived the encounter), completely oblivious to the new variable.

But how do you make it regonise new shapes? It has a video thing to recognise new shapes, and will be encouraged to investigate them because it needs power, so we are looking at artificial intelligence that is made to survive, as that is what intelligence is - the seeking of survival. If this thing seeks survival it has intelligence.

Seeking survival has nothing to do with intelligence, I've been over and have modified the lines of BASIC that control my little robot, adaptability to a situation or variable is all in the hands of the programmer.

I have the feeling I completely missed the point of your post, so please correct me if I'm wrong.

[harrythehorton]

Artificial in this subject can be defined as:
1. made by human skill; produced by humans (opposed to natural): artificial flowers.
2. imitation; simulated; sham: artificial vanilla flavoring.

Intelligence:
1. capacity for learning, reasoning, understanding, and similar forms of mental activity; aptitude in grasping truths, relationships, facts, meanings, etc.

Rather than debating definitions, let's have them defined as they are generally understood with the subject.

First off. Never say never. Don't say something could never be possible, because honestly, such an assumption will in many cases be proven wrong with time(granted it's somewhere near semi-understandable or conceptable bounds).

Artificial Intelligence: a simulation of a capacity for learning, reasoning, understanding, and similar forms of mental activity(to combine definitions).

Now with the literal definition of the terms. Yes, it's possible. It's being done now, there are too many examples that are meant to SIMulate intelligence. Though, that's not the definition people are normally referring to when talking about the future.

So, the question is: are sentience, comprehension, reasoning, and a valid consciousness possible within the bounds of a machine?

Many people have tried to write a program that reacts to user input(variables), who's responses are only based off the program's functions (predetermined, in a sense, responses... even if set to random[random what =)]). This is only a simulation. Then some programs were written which could connect, only mathematically, relationships between words and how they relate to each other saying "IT UNDERSTANDS!!!". How can something have an understanding of terms or even definitions with nothing to relate it to physically?

A program doesn't 'understand' it only reacts to variables based on predetermined responses. Even if your game AI choses a path through some buildings at random, to get to a target, it's still predetermined. There is code there that inables the representation of a character to "move" randomly. Random is part of a function.

So when is sentience possible?
1. let's make sure we're out of the confines of only a coded file.
2. let's embed whatever code we've written into something physical. Robotics come to mind, neh? (Catch that anyone ^.^?)E.G.

In order to be sentient, something needs to be able to precieve. without perceptions of reality, there's really nothing to truly experience, therefore no understanding. . . without understanding how is there intelligence.

Experience is the base of everything that we are. Experience is what urges us to act and react the way that we do, sorta like programming. Only these reactions wern't predetermined at birth. Were they? no.
Sure we make choices of whether or not to do something, but to even understand what that something is, we need reality experience. . . otherwise how would there be sentience.

We experience through action, and through our mental representation of an action(normally dependent one what we've heard or seen). So, a machine would need to be the same way.

Okay so we need something that can physically react with society. Robotics covers that. There are robots than can to an extent learn... memory and experience, check.

We don't have a readout/access to all of the happenings in our body. Sure there are some responses we get through feeling, but we don't calculate the information traveling through our minds until it gets to our conscious perception do we? With that said. . . robotics would need the same push in their programming to achieve true sentience.

A program needs to be semi cutoff from the base functions, which keep it going. There should be something in the program that receives input on a constant timescale, with no connection to the electronic support functions. This area of the program would need to project two images. Visuals, and the best representation of the mind's eye (aka thought). all processing known to this 'conscious' section of the program would be done within this mind's eye through a mixture of images, sounds, and how reactions are played out. Basically a three dimensional virtual reality would be built in order for the consciousness to perceive things as similar to us as possible.(this get's more complex, but it's difficult to explain)

Only a theory, but consciousness is possibly a check. That coupled with the ability for robotics to learn through visuals and reactions to their own physical action. Well that's so many steps closer to the goal of true intelligence through an artificial system.

And that is what we're talking about here 'intelligence through an artificial system'. That is the meaning that generally comes to thought(due to a process of retrieving emotions, sights, sounds, feelings related to the term artificial intelligence). Yes you could look at the term literally, but don't act like we do that with everything[ex. 'a lot...' it's casually used to refer to a large number of the subject. . . not an actual LOT of whatever.. yeh that's the root, but we're not thinking of it literally]

Conclusion. Yes, i believe artificial intelligence is possible, granted:
1. Program is embedded within a physical robotic machine
2. the sensory input is relayed to the 'conscious' part of the program
3. the conscious is connected to memory through will of recall and situational necessity/happenings
4. The consciousness is separated from the base electronic stabilizing functions
5. Through the consciousness has the ability to reason using the memories gathered purely through previous sensory input and actions.

The question from me now is... can a machine have will?

In order to bring things closer and closer to reality, the trick is not writing a function to handle a full action to represent something... it's separating each and every possible element of anything done in an organism and representing the most BASIC parts of each action to the most petite possible detail. So, this new list of extremely basic functions have to work together, yet separate in their own way, to present each and every action preformed by a machine.

At that point, we can only hope sentience is born.

[Rainbow]

The device is made of a compound known as duroquinone. This molecule resembles a hexagonal plate with four cones linked to it, "like a small car," explained researcher Anirban Bandyopadhyay, an artificial intelligence and molecular electronics scientist at the National Institute for Materials Science at Tsukuba in Japan. Duroquinone is less than a nanometer, or a billionth of a meter large. This makes it hundreds of times smaller than a wavelength of visible light. The machine is made of 17 duroquinone molecules. One molecule sits at the center of a ring formed by the remaining 16. The entire invention sits on a surface of gold.

This is the latest development concerning AI technology.
Link : Tiny Brain-Like Computer Created | LiveScience