Artificial Intelligence? Why not Real Intelligence?

[tar]

A human is in and of reality.

We experience a "here and now".

We internalize the outside world and build an analog model of it.

When our model fits the actual world that we have a model of, it makes us feel good.

When the actual world does not fit our model, we adjust our model to fit, or when within our power, we adjust the world to fit.

When things do not fit, it makes us feel bad.

 

We have hunger, thirst, and other "feelings" of emptiness or lack that "need" to be fulfilled for things to be right.

 

We are born, we grow, we learn about the world, we explore the world, we share our experience of it with others, and try to fix what is "wrong" and enjoy what is right. And we eventually die.

 

If we were to give a man made machine, analogous facilities to ours, and it was able to actually sense the world and move around in it and build an internal model of it AND we would think of a way to reward or pleasure the machine when "fits" between the internal model and the external world were found, and a way to punish or pain the machine when not...then the machine could learn about the world, by itself, and its knowledge would be actual knowledge of the world, not subject to the programmer's code. Efforts that the machine made, would be on its own behalf. It would not be artificial. It's senses would internalize the external real world and build an analog model of it, and it would have reason to explore and learn and fill in any "gaps" between its model and the world it was of and in.

 

Regards, TAR2

Edited October 1, 2011 by tar

[Schrödinger's hat]

A human is in and of reality.

We experience a "here and now".

We internalize the outside world and build an analog model of it.

When our model fits the actual world that we have a model of, it makes us feel good.

When the actual world does not fit our model, we adjust our model to fit, or when within our power, we adjust the world to fit.

When things do not fit, it makes us feel bad.

 

We have hunger, thirst, and other "feelings" of emptiness or lack that "need" to be fulfilled for things to be right.

 

We are born, we grow, we learn about the world, we explore the world, we share our experience of it with others, and try to fix what is "wrong" and enjoy what is right. And we eventually die.

 

If we were to give a man made machine, analogous facilities to ours, and it was able to actually sense the world and move around in it and build an internal model of it AND we would think of a way to reward or pleasure the machine when "fits" between the internal model and the external world were found, and a way to punish or pain the machine when not...then the machine could learn about the world, by itself, and its knowledge would be actual knowledge of the world, not subject to the programmer's code. Efforts that the machine made, would be on its own behalf. It would not be artificial. It's senses would internalize the external real world and build an analog model of it, and it would have reason to explore and learn and fill in any "gaps" between its model and the world it was of and in.

 

(from wiktionary) Artificial:

Etymology

Via Old French (French: artificiel), from Latin artificialis from artificium "skill", from artifex, from ars "skill", and -fex, from facere "to make".

 

When I read artificial intelligence I think intelligence constructed with skill, rather than fake intelligence.

If we were to create something that is indistinguishable from intelligence (score on the turing test would be one metric for this, although probably not the only thing required), then I call it intelligent.

[insane_alien]

yeah, the prefix artificial really just describes that the intelligence(which would be real intelligence) arises through constructed means rather than being an emergent property of life.

 

fake intelligence could be applied to the artificial 'intelligence' present in modern computer games. where the intelligence is not actually very high. It is however designed to look intelligent while actually being pretty minimalist and mechanical. basically, the AI opponents in games like Modern Warfare 2 are faking being smart by having lots of scripted actions with a relatively(compared to true intelligence) simple decision tree to choose which to perform and when to perform it.

 

of course, if you go far enough down that route to the point where you can fake true intelligence in every aspect then you have created true artificial intelligence. Current best guess is that it is easier to approach the problem by attempting to create a true artifical intelligence from the start rather than a limited system.

[Schrödinger's hat]

fake intelligence could be applied to the artificial 'intelligence' present in modern computer games. where the intelligence is not actually very high. It is however designed to look intelligent while actually being pretty minimalist and mechanical. basically, the AI opponents in games like Modern Warfare 2 are faking being smart by having lots of scripted actions with a relatively(compared to true intelligence) simple decision tree to choose which to perform and when to perform it.

 

Indeed. Intelligence also don't have to be artificial (made by man) to be this fake kind of intelligence.

The behavior of some insects comes to mind. Primarily the digger wasp.

[tar]

Schrödinger's hat,

 

Interesting the the Turing test involves human language. Along with my thought that "language is communication between two minds that already know each other" it sort of skews any determination a human would make.

 

It is, the Turing test that is, really testing "how well does this mind know me, and how well do I know this mind?" If it appears that the mind under test knows you well, and you know the mind well, then it passes the test. But it is the mind of the programmers who you are really communicating with. So yes, the machine that passes the test "has" human intelligence, but it is an extension of the programmers' wit, and it is the programmers' wit you are testing, and trying to outsmart. If you cannot determine, or judge whether you are communicating with a machine or a human, it is because in both cases, you ARE communicating with a human mind.

 

To test the Digger Wasp intelligence, you would have to already know the mind of the Digger Wasp. We don't know what the Wasp is checking for when it enters the nest again after realigning the prey at the mouth of the hole. Perhaps it has just determined the size and shape and dragability and wheight distribution of the prey and with that information, is "rehearsing" the route into the whole, and checking for a route and storage space that will work.

 

Not unlike a golfer taking a few practice swings a foot away from the ball before stepping up to the putt. If we would distract the golfer, and move the ball, he/she would have to go through the whole routine again.

 

Regards, TAR2

 

anyway, there is something wrong with a cricket with no antennae. Even a Digger Wasp knows that!

 

insane_alien,

 

So if artificial is "skilled" and "faked" is mimicked or borrowed, then "true" intelligence would be that which "emerges" as an entity aquires a skill, on its own, and retains it, and uses it on its own behalf.

 

Live things have a leg up on machines, in that they have aquired skills and retained them as real components of their being, over many millions of years of interaction with the environment.

 

Schrödinger's hat,

 

Would you say a cherry tree that flowers in the spring, is skilled? I can't do that. All I have is human skills.

 

Regards, TAR2

[Schrödinger's hat]

Schrödinger's hat,

 

Interesting the the Turing test involves human language. Along with my thought that "language is communication between two minds that already know each other" it sort of skews any determination a human would make.

 

It is, the Turing test that is, really testing "how well does this mind know me, and how well do I know this mind?" If it appears that the mind under test knows you well, and you know the mind well, then it passes the test. But it is the mind of the programmers who you are really communicating with. So yes, the machine that passes the test "has" human intelligence, but it is an extension of the programmers' wit, and it is the programmers' wit you are testing, and trying to outsmart. If you cannot determine, or judge whether you are communicating with a machine or a human, it is because in both cases, you ARE communicating with a human mind.

 

To test the Digger Wasp intelligence, you would have to already know the mind of the Digger Wasp. We don't know what the Wasp is checking for when it enters the nest again after realigning the prey at the mouth of the hole. Perhaps it has just determined the size and shape and dragability and wheight distribution of the prey and with that information, is "rehearsing" the route into the whole, and checking for a route and storage space that will work.

 

Not unlike a golfer taking a few practice swings a foot away from the ball before stepping up to the putt. If we would distract the golfer, and move the ball, he/she would have to go through the whole routine again.

Hmm, but if you were to keep moving the ball, the golfer would quickly get irate and stop you. On top of this, we know of systems that act in complicated but completely deterministic ways by a set of rigidly defined rules.

 

Perhaps the turing test was too specific. What I had in mind when I said that was a more general version where one could interact with the entity in a variety of ways. If, after that interaction, you cannot distinguish between the entity you interacted with, and an entity that you consider intelligent. Then for the purposes of that type of interaction you would consider that entity intelligent.

 

Would you say a cherry tree that flowers in the spring, is skilled? I can't do that. All I have is human skills.

Hard to say.

 

If the cherry tree were to suddenly make an entity that acted intelligently then I would call that entity an artificial intelligence.

If it were to build a house, I'd call it an artificial shelter.

 

I do not know where to draw the line, but the blossoms are more easily explainable through known laws and interactions than my ability to make a toy out of wood for my nephew. I don't consider my ability to grow hair or fingernails a skill. There is something difficult to define, but important about this distinction.

 

Perhaps this definition for 'a skill' might work:

An ability which can be absent in one entity/place, and then transmitted/induced from/by another place/entity without exchange of matter.

[Mrs Zeta]

One could also consider the possibility that if a machine will ever be able to reach human intelligence and any other qualities that make us human, then it would have to mimic a human both in structure and organisation to such a depth and to such a degree that it will become one and the same to its biological counterpart (matching our biological capabilities, frailties and shorcomings). So why bother in the first place?

[Schrödinger's hat]

One could also consider the possibility that if a machine will ever be able to reach human intelligence and any other qualities that make us human, then it would have to mimic a human both in structure and organisation to such a depth and to such a degree that it will become one and the same to its biological counterpart (matching our biological capabilities, frailties and shorcomings). So why bother in the first place?

 

Well if we have the blueprints we could tweak it a little.

Or if it's on a computer we can watch what's happening in greater detail than we possible could in squishy-land.

That way we could learn a lot more about it/ourselves.

 

Of course when it comes to modifications there are ethical concerns to worry about.

[tar]

Mrs Zeta,

 

Yeah, if you want to create an entity with human intelligence, get a young woman and a man together in a private place with a bottle of wine and no birth control. Do this every night for about a month, and 9 or 10 months later... wah la.

 

Regards, TAR2

 

Schrödinger's hat,

 

Well I get your distinction, and know that there is "something" about what we do that is different than what a tree does. But I think it more of a distinction in degree, than a distinction in kind.

 

Life on this planet is interwoven. Some of the chemicals we need to survive are produced by other living things. The amount of available chemicals that we use to grow and repair and metabolize in the biosphere is a result of living things living and dying and being used by other living things. For 100s of billions of years. We "fit" the planet. Not only our evolutionary history, but that of countless other species is involved.

 

Like a fish "conditioning" the tank.

 

If we know how to metabolize, and grow, and differenciate organs and such, it has something to do with the "skill" of the mitochondria, and the DNA and RNA that pass a "working" pattern to the next generation.

 

Our brains have analogies, sometimes close analogies to other mammals.

 

I would not hastily pooh pooh the accomplishments and skill of a Digger Wasp and religate their existence to a "deterministic system" with no intelligence involved. We are just better at doing what we do, and the wasp better at doing what it does. And the cherry tree better at being a cherry tree.

 

The systems are the same in many regards. And at no point do emergent properties emerge, without something to emerge from.

 

If you roll the dice you get 2,3,4,5,6,7,8,9,10,11 or 12.

 

You will never roll a Queen of Spades. For that you need a deck of cards.

 

Regards, TAR2

 

Quick thought on the Digger Wasp not dragging an antenaeless cricket.

 

I have had a number of occassions where I chopped up a fallen large branch or tree into draggable components. I learned to always pick up the cut and drag the branch that way. You have the heaviest part in your hand and the branches which usually branch away from the trunk slide nicely over the ground and you and your load are the "thinnest" possible profile for passing between trees and other obstacles. In fact, in this orientation the branches "yield" to the obstacles. If you grab one of the small branches and drag your load sideways there is MUCH more resistance, sometimes impossible resistance. And forget trying to drag a branch from the tip.

 

If you were dragging a branch into a hole, you would do it with a grip on the cut. So would a Digger Wasp drag a cricket.

[questionposter]

I've thought a lot about how intelligence would work in a complex robot, and I just can't see a difference between what it's doing and what many living things do. While it may have the ability to chose, how it chose is the same of how the subconscious of other living things make decisions based on how it plays out in my mind. It may be that "artificial" intelligence has been invented 4.5 billion years ago.

Edited October 19, 2011 by questionposter

[edwardreed]

What is Artificial Intelligence. I would say we have it already.

 

!

Moderator Note

 

Answer plagiarized, so it has been deleted. To see it go to the link

http://answers.yahoo.com/question/index?qid=20110307112440AAsJ66L

[tar]

edwardreed,

 

Yes we have a human "bias". We have to, that is what we are. My thinking is along your line, that intelligence is best considered as something that everything has to some degree, that human's have a whole lot of. We have, through the mechanisms that evolution has developed for us, and through exploration and experiment, discovered much about the world we are part of. Other things do not seem to be so aware and capable.

 

How we do it, how we "emerge" with this awareness is my facination. I have an overall goal, in my investigations over the past year, to understand the "meaning" behind language. The universal grammar that infants are in possession of, does in my guess, have everything to do with what a human is, and does. It is not something magically gifted from above, but something ACTUALLY developed from millions of years of interaction with reality, and "remembered" through the passing on of the resulting genetic patterns to the next generation.

 

In this, our intelligence is not wholly our own, but a reflection of the world from which we emerged, and in which we reside.

 

Regards, TAR2

[kitkat]

edwardreed,

 

Yes we have a human "bias". We have to, that is what we are. My thinking is along your line, that intelligence is best considered as something that everything has to some degree, that human's have a whole lot of. We have, through the mechanisms that evolution has developed for us, and through exploration and experiment, discovered much about the world we are part of. Other things do not seem to be so aware and capable.

 

How we do it, how we "emerge" with this awareness is my facination. I have an overall goal, in my investigations over the past year, to understand the "meaning" behind language. The universal grammar that infants are in possession of, does in my guess, have everything to do with what a human is, and does. It is not something magically gifted from above, but something ACTUALLY developed from millions of years of interaction with reality, and "remembered" through the passing on of the resulting genetic patterns to the next generation.

 

In this, our intelligence is not wholly our own, but a reflection of the world from which we emerged, and in which we reside.

 

Regards, TAR2

 

Your comment that we have human "bias". We have to, that is what we are. I disagree, science has come a long way in regards to studying other species behavior so I would think that we would realize that defining intelligence based on human ananomical abilities and ignoring other species ananomical abilities to set the standards for defining intelligence is illogical. Defining intelligence is unique to every species and it should take into consideration their morphology and the extent of their sensory abilities that allows all of them to see the world differently from us and unique to them. Processing information from the environment is not the same process in every species and we as humans are smart enough now to realize that.

[tar]

kitkat,

 

I do not disagree with what you say. Merely stating that we have to have a human bias, because there is not "another" perspective we can take. We can put ourselves in the shoes of another entity, but it is always "us" we are putting in the "others" shoes.

 

Are you saying one can reach a level of understanding that is not a human one?

 

Regards, TAR2

[kitkat]

kitkat,

 

I do not disagree with what you say. Merely stating that we have to have a human bias, because there is not "another" perspective we can take. We can put ourselves in the shoes of another entity, but it is always "us" we are putting in the "others" shoes.

 

Are you saying one can reach a level of understanding that is not a human one?

 

Regards, TAR2

 

That is a very good question, however, the level of understanding between a human and another mammal is not so vast that we cannot communicate at all. We share many of the same principles of sexual behavior, nurturing offspring, emotional responses and so on. When you say, "but it is always "us" we are putting in the other's shoes" is not all together correct. Empathy is mirroring neurons from one entity to the one that is effected by it and this appears almost universal in response to most creatures we share this planet with.

 

We understand when a snake uses it rattle when we are too close to it that it means to back off we are in its space. People use verbal or body language to create the same message. The snake cannot speak our language obviously but we still understand what the snake is telling us. It is the same when you are in a foreign country trying to communicate to the natives by using your language. You end up using body language in gestures to hopefully get them to understand what you need.

 

This is no different from what other mammals use in their abilities to be understood when interacting with another type of creature. Cats and dogs are always invented new ways to get their point across and when we feel we have trained them but in reality many of them have trained us.