Extreme Artificial Intelligence
-
Hey guys & ladies (to be gender insensitive), a theoretical thought, if a computer program simulates the human brain very accurately, does that make the program self - aware?
I don't think there's anything extreme about the prospect. What does it really mean for a program to be self-aware? If we can say a program can be "aware" of *anything* then surely if it's model in any way includes any aspect of itself, then it IS self-aware. Many people probably will have a hard time accepting this. But to those I ask them to specify by what criterion it is we can say that the brain is self-aware. And was the brain self-aware before we knew we have brains? I'm not just speaking of an evolutionary perspective here, but each and every one of us were born having not the slightest clue that we possessed a brain - it is something we became aware of years into our lives. In fact, we weren't even aware of our own individuality (which of course what we mean when we say a human is self-aware, rather than the more specific proposal that the brain is self-aware) until years into our lives. And no matter how wise and old you may be, you are *still* unaware of the vast majority of what is going on inside your head. When you speak, you are merely aware of part of the "top layer" occupied with expressing some idea or opinion, while all the lower-level processing required to unfold words into phonems and phonems into sequences of precise motor action (and surely a lot I don't know about) is totally transparent. In the end, I think this whole issue of self-awareness is just a special case of the larger problem of perception. We all experience it and so can agree that it is a real phenomenon. There has never been a hint of solid evidence it is anything but a "side effect" of the physical activity in our brain, but nor do we have even a beginning of a clue of how and why the phenomena we observe in the brain actually lead to perception. In other words, if we were faced with a machine other than a brain that implemented self-awareness in a different way (if that is even possible) we have no reason at all to believe we would recognize that is was - at least not unless the machine could somehow express this self-awareness in a manner detectable and understandable to us. Numenta are seemingly having some success in building "intelligent" machines, or at least machines that are cracking the kind of problems 50 years of traditional AI research could not. And with some "animal-like" characteristics, such as being much better at recognizing visual objects when they move and more generally a dependency on the temporal aspect in sensing that has so often been ignored. :)
-
I don't think there's anything extreme about the prospect. What does it really mean for a program to be self-aware? If we can say a program can be "aware" of *anything* then surely if it's model in any way includes any aspect of itself, then it IS self-aware. Many people probably will have a hard time accepting this. But to those I ask them to specify by what criterion it is we can say that the brain is self-aware. And was the brain self-aware before we knew we have brains? I'm not just speaking of an evolutionary perspective here, but each and every one of us were born having not the slightest clue that we possessed a brain - it is something we became aware of years into our lives. In fact, we weren't even aware of our own individuality (which of course what we mean when we say a human is self-aware, rather than the more specific proposal that the brain is self-aware) until years into our lives. And no matter how wise and old you may be, you are *still* unaware of the vast majority of what is going on inside your head. When you speak, you are merely aware of part of the "top layer" occupied with expressing some idea or opinion, while all the lower-level processing required to unfold words into phonems and phonems into sequences of precise motor action (and surely a lot I don't know about) is totally transparent. In the end, I think this whole issue of self-awareness is just a special case of the larger problem of perception. We all experience it and so can agree that it is a real phenomenon. There has never been a hint of solid evidence it is anything but a "side effect" of the physical activity in our brain, but nor do we have even a beginning of a clue of how and why the phenomena we observe in the brain actually lead to perception. In other words, if we were faced with a machine other than a brain that implemented self-awareness in a different way (if that is even possible) we have no reason at all to believe we would recognize that is was - at least not unless the machine could somehow express this self-awareness in a manner detectable and understandable to us. Numenta are seemingly having some success in building "intelligent" machines, or at least machines that are cracking the kind of problems 50 years of traditional AI research could not. And with some "animal-like" characteristics, such as being much better at recognizing visual objects when they move and more generally a dependency on the temporal aspect in sensing that has so often been ignored. :)
dojohansen wrote:
I don't think there's anything extreme about the prospect. What does it really mean for a program to be self-aware?
The extreme part comes in because I think that if a program had to mimic human intelligence, at least at a level were one can have a meaningful conversation with the program and that the program be environment aware, then that's extreme artificial intelligence, I think self awareness stems from intelligence which most animals possess but current programs seem to fall short of. Having a program able to demonstrate such attributes can be considered extreme artificial intelligence, more like the peak of AI.:cool: But i do agree with you on the other points you stipulated.:thumbsup:
“Be at war with your vices, at peace with your neighbors, and let every new year find you a better man or woman.”
-
dojohansen wrote:
I don't think there's anything extreme about the prospect. What does it really mean for a program to be self-aware?
The extreme part comes in because I think that if a program had to mimic human intelligence, at least at a level were one can have a meaningful conversation with the program and that the program be environment aware, then that's extreme artificial intelligence, I think self awareness stems from intelligence which most animals possess but current programs seem to fall short of. Having a program able to demonstrate such attributes can be considered extreme artificial intelligence, more like the peak of AI.:cool: But i do agree with you on the other points you stipulated.:thumbsup:
“Be at war with your vices, at peace with your neighbors, and let every new year find you a better man or woman.”
Boringly, we probably agree on all points then. A program demonstrating anything resembling generalized intelligence at anywhere near a human capability level would indeed be extreme. I just don't think being self-aware necessarily has that much to do with intelligence, even if it did emerge in us that way. Self-modifying code is in a (limited) sense self-aware. :)
-
Boringly, we probably agree on all points then. A program demonstrating anything resembling generalized intelligence at anywhere near a human capability level would indeed be extreme. I just don't think being self-aware necessarily has that much to do with intelligence, even if it did emerge in us that way. Self-modifying code is in a (limited) sense self-aware. :)
dojohansen wrote:
I just don't think being self-aware necessarily has that much to do with intelligence, even if it did emerge in us that way
I think for a program to induce that it is self aware, then it must have a "mental capacity" to do just that. Induction at such levels requires intelligence, for you to induce that you exist and therefore self aware requires some thought and intelligence.
dojohansen wrote:
Self-modifying code is in a (limited) sense self-aware
Ability to modify it's own coding is another "mentally" demanding activity without a form of intelligence i doubt that that program can be able to modify it's on coding. Therefore self awareness requires some intelligence at a certain level. :)
“Be at war with your vices, at peace with your neighbors, and let every new year find you a better man or woman.”
-
currently it has not been achieved yet, but are programs that use sensory processing such as computer vision systems self aware?
No.
-
Hey guys & ladies (to be gender insensitive), a theoretical thought, if a computer program simulates the human brain very accurately, does that make the program self - aware?
Not really. This "code" is kinda the same for the robot and a person:
if self is touching("Hot Stove") feelHot();
But for the person, the function feelHot will have several million, if not billion lines of code. Robot will have much much much less. In the function, the robot will probably have something that tells him to move away his hand and maybe inspect it for any damage, but that's it. The person, on the other, would actually feel the pain. Maybe they will start crying, maybe they will get ice or maybe do something else. It is impossible to predict. If you write a new genius sorting algorithm computer will not be aware of it. Other people may inspect it and find it really smart and cool but computer will just execute it faster than a slower algorithm. You can teach a 10 year old the basic math and he will probably be able to use the algorithm, and perform the operations specified there, but he will probably not be aware of how it works. Even the basic functions for human, have unimaginable complexity. For computer adding 2 numbers is just adding 2 numbers. But for humans, each number can also invoke specific memory, feeling, etc.
-
No.
Well but at least they are aware of what they recognize. :laugh:
“Be at war with your vices, at peace with your neighbors, and let every new year find you a better man or woman.”
-
Not really. This "code" is kinda the same for the robot and a person:
if self is touching("Hot Stove") feelHot();
But for the person, the function feelHot will have several million, if not billion lines of code. Robot will have much much much less. In the function, the robot will probably have something that tells him to move away his hand and maybe inspect it for any damage, but that's it. The person, on the other, would actually feel the pain. Maybe they will start crying, maybe they will get ice or maybe do something else. It is impossible to predict. If you write a new genius sorting algorithm computer will not be aware of it. Other people may inspect it and find it really smart and cool but computer will just execute it faster than a slower algorithm. You can teach a 10 year old the basic math and he will probably be able to use the algorithm, and perform the operations specified there, but he will probably not be aware of how it works. Even the basic functions for human, have unimaginable complexity. For computer adding 2 numbers is just adding 2 numbers. But for humans, each number can also invoke specific memory, feeling, etc.
Complexity does not imply self awareness. The ability to be unpredictable does not imply self awareness, but the ability to associate stimuli to a specific response and the ability to self monitor imply self awareness, there is no need to be as complex as a human to attain self awareness. Or and it's not the computer in question but the computer program. :)
“Be at war with your vices, at peace with your neighbors, and let every new year find you a better man or woman.”
-
I was not sure the moment i asked the question, but after thinking about it and taking into consideration views from others, my view evolved and became more inclined to the fact that some programs can be or are already self aware. :)
“Be at war with your vices, at peace with your neighbors, and let every new year find you a better man or woman.”
BupeChombaDerrick wrote:
the fact that some programs can be or are already self aware
Such as? Name a few...
"The clue train passed his station without stopping." - John Simmons / outlaw programmer "Real programmers just throw a bunch of 1s and 0s at the computer to see what sticks" - Pete O'Hanlon "Not only do you continue to babble nonsense, you can't even correctly remember the nonsense you babbled just minutes ago." - Rob Graham
-
BupeChombaDerrick wrote:
the fact that some programs can be or are already self aware
Such as? Name a few...
"The clue train passed his station without stopping." - John Simmons / outlaw programmer "Real programmers just throw a bunch of 1s and 0s at the computer to see what sticks" - Pete O'Hanlon "Not only do you continue to babble nonsense, you can't even correctly remember the nonsense you babbled just minutes ago." - Rob Graham
My view is that awareness is mostly about associating a stimuli to a particular response. At least any program that associates a keyboard input to a particular response is aware of those key strokes. So a word processor is aware if a keyboard button is pressed, but to be self aware a program only needs to have some self monitoring capability such as a feedback about it's performance. :)
“Be at war with your vices, at peace with your neighbors, and let every new year find you a better man or woman.”
-
My view is that awareness is mostly about associating a stimuli to a particular response. At least any program that associates a keyboard input to a particular response is aware of those key strokes. So a word processor is aware if a keyboard button is pressed, but to be self aware a program only needs to have some self monitoring capability such as a feedback about it's performance. :)
“Be at war with your vices, at peace with your neighbors, and let every new year find you a better man or woman.”
BupeChombaDerrick wrote:
a word processor is aware if a keyboard button is pressed
Not really in terms of "self aware", but it is "aware" of key presses when the system signals that a key press has been done. Two completely different definitions. It just sits there idling waiting for a keyboard press event to be triggered. My idea of "self awareness" is one in which something can take care of itself without necessarily having interaction with something else. A word processor cannot and will not do anything until the user interacts with it.
BupeChombaDerrick wrote:
to be self aware a program only needs to have some self monitoring capability
Which still boils down to someone coding that capability. Looking through this entire thread, there are a lot of interesting points made by everyone...
"The clue train passed his station without stopping." - John Simmons / outlaw programmer "Real programmers just throw a bunch of 1s and 0s at the computer to see what sticks" - Pete O'Hanlon "Not only do you continue to babble nonsense, you can't even correctly remember the nonsense you babbled just minutes ago." - Rob Graham
-
currently it has not been achieved yet, but are programs that use sensory processing such as computer vision systems self aware?
-
BupeChombaDerrick wrote:
a word processor is aware if a keyboard button is pressed
Not really in terms of "self aware", but it is "aware" of key presses when the system signals that a key press has been done. Two completely different definitions. It just sits there idling waiting for a keyboard press event to be triggered. My idea of "self awareness" is one in which something can take care of itself without necessarily having interaction with something else. A word processor cannot and will not do anything until the user interacts with it.
BupeChombaDerrick wrote:
to be self aware a program only needs to have some self monitoring capability
Which still boils down to someone coding that capability. Looking through this entire thread, there are a lot of interesting points made by everyone...
"The clue train passed his station without stopping." - John Simmons / outlaw programmer "Real programmers just throw a bunch of 1s and 0s at the computer to see what sticks" - Pete O'Hanlon "Not only do you continue to babble nonsense, you can't even correctly remember the nonsense you babbled just minutes ago." - Rob Graham
Paul Conrad wrote:
My idea of "self awareness" is one in which something can take care of itself without necessarily having interaction with something else.
But we interact with our environment and our own internal states thus we respond to environmental stimulation as well as to feedback from some of our internal mechanisms. Without this interaction we can be completely oblivious of our own existence.
Paul Conrad wrote:
Which still boils down to someone coding that capability.
Even humans are hardwired to be self aware following some form of design. :)
“Be at war with your vices, at peace with your neighbors, and let every new year find you a better man or woman.”
-
Paul Conrad wrote:
My idea of "self awareness" is one in which something can take care of itself without necessarily having interaction with something else.
But we interact with our environment and our own internal states thus we respond to environmental stimulation as well as to feedback from some of our internal mechanisms. Without this interaction we can be completely oblivious of our own existence.
Paul Conrad wrote:
Which still boils down to someone coding that capability.
Even humans are hardwired to be self aware following some form of design. :)
“Be at war with your vices, at peace with your neighbors, and let every new year find you a better man or woman.”
Cogito ergo sum ("I think, therefore I am") proposed by René Descartes.... I know this will probably really open it up some...
"The clue train passed his station without stopping." - John Simmons / outlaw programmer "Real programmers just throw a bunch of 1s and 0s at the computer to see what sticks" - Pete O'Hanlon "Not only do you continue to babble nonsense, you can't even correctly remember the nonsense you babbled just minutes ago." - Rob Graham
-
Apparently so: Robot sees itself in a mirror[^]
Nice robot, any further advancement in this field will definitely lead to a completely independent robot (of course not one equipped with a deadly machine gun) that can go around and learn about it's environment. :) Nice link:thumbsup:
“Be at war with your vices, at peace with your neighbors, and let every new year find you a better man or woman.”
-
Paul Conrad wrote:
My idea of "self awareness" is one in which something can take care of itself without necessarily having interaction with something else.
But we interact with our environment and our own internal states thus we respond to environmental stimulation as well as to feedback from some of our internal mechanisms. Without this interaction we can be completely oblivious of our own existence.
Paul Conrad wrote:
Which still boils down to someone coding that capability.
Even humans are hardwired to be self aware following some form of design. :)
“Be at war with your vices, at peace with your neighbors, and let every new year find you a better man or woman.”
Well it seems you're already pretty convinced of what "self awareness" is (you defined it for us in an earlier post), so I won't comment on that. What I found most compelling is your signature: "...every year find you a better .. woman." SWEET! A better (new?) woman every year! I might actually enjo NO WAIT, HONEY, I DIDN'T WRITE THAT. PUT THE BAT DOWN! OW. OW. OW.
David --------- Empirical studies indicate that 20% of the people drink 80% of the beer. With C++ developers, the rule is that 80% of the developers understand at most 20% of the language. It is not the same 20% for different people, so don't count on them to understand each other's code. http://yosefk.com/c++fqa/picture.html#fqa-6.6 ---------
-
Well it seems you're already pretty convinced of what "self awareness" is (you defined it for us in an earlier post), so I won't comment on that. What I found most compelling is your signature: "...every year find you a better .. woman." SWEET! A better (new?) woman every year! I might actually enjo NO WAIT, HONEY, I DIDN'T WRITE THAT. PUT THE BAT DOWN! OW. OW. OW.
David --------- Empirical studies indicate that 20% of the people drink 80% of the beer. With C++ developers, the rule is that 80% of the developers understand at most 20% of the language. It is not the same 20% for different people, so don't count on them to understand each other's code. http://yosefk.com/c++fqa/picture.html#fqa-6.6 ---------
The "better man or woman" part is merely to be gender insensitive not that I'am a woman, I am a man. Just want to make women feel free to see that they are being considered in the statement. :laugh: so you can only "HONEY" me if you were a woman and I can only "HONEY" a woman. ;)
“Be at war with your vices, at peace with your neighbors, and let every new year find you a better man or woman.”
-
Our intelligence is not a result of computations. We are consummate pattern-matchers. Here is a really simplified version of how it goes: When we perceive something, it causes a certain bunch of sensory neurons to fire, which correspond directly to that perception. The neurons connected to those sensory neurons fire in turn if they recognize a pattern there -- for example, some neurons only fire if they see a vertical bar traveling from left to right, or other specific patterns like that. Then the next level of connected neurons fire if they recognize a particular pattern in the level before them, and so forth. We learn by building up patterns of patterns. The match to a pattern pops up automatically, or in other words, perceiving and recalling a matching previous pattern happen because the perception and the recall are linked by sharing the same set of neurons in the middle. For an example of how this works, take driving. When you first got behind the wheel as a kid, everything seemed very unfamiliar. All the knobs were confusing, and you probably had to concentrate to remember which pedal was which. You probably had trouble recognizing following distances and when to turn to fit into a parking space and that kind of thing. But with practice, your brain began to recognize and store the patterns of driving, until almost all of driving became subconscious pattern-matching -- the lines on the road should be at particular distances, the feel of the brake matches to how quickly or how slowly the car comes to a stop, et cetera -- we don't have to think about any of these things because they match stored patterns in our minds. We don't have to consciously think about anything unless it breaks our expectations. Unexpected or unknown things draw our attention because they defy the patterns we know. In contrast, a computer is terrible at pattern-matching. Many, many man-years went into the Google search algorithm, but really, what it's doing is trying to mimic the natural human ability to glance over a list and recognize what you are looking for out of it. This very basic ability has to be painstakingly coded into the computer. If you lined up a bunch of toys and asked a preschooler to hand you the "meanest one," the preschooler will be able to match his or her idea of "meanness" to the various traits of the toys and decide which one is the most mean. The computer, on the other hand, has no ability to take the concept of "mean" and expand it to apply to a toy, *unless a human writes an algorithm d
That is by far the best description of AI that i have heard. I think you are totally right in saying that the approach must be different. In fact, the result will not be a computer at all, since a computer does what you tell it to. It is not the case that a self-aware entity will do as you ask. If i ask a lion to say cheese, it will most likely eat me instead.
-
BupeChombaDerrick wrote:
a word processor is aware if a keyboard button is pressed
Not really in terms of "self aware", but it is "aware" of key presses when the system signals that a key press has been done. Two completely different definitions. It just sits there idling waiting for a keyboard press event to be triggered. My idea of "self awareness" is one in which something can take care of itself without necessarily having interaction with something else. A word processor cannot and will not do anything until the user interacts with it.
BupeChombaDerrick wrote:
to be self aware a program only needs to have some self monitoring capability
Which still boils down to someone coding that capability. Looking through this entire thread, there are a lot of interesting points made by everyone...
"The clue train passed his station without stopping." - John Simmons / outlaw programmer "Real programmers just throw a bunch of 1s and 0s at the computer to see what sticks" - Pete O'Hanlon "Not only do you continue to babble nonsense, you can't even correctly remember the nonsense you babbled just minutes ago." - Rob Graham
Your "Self Awareness" sounds to me more like "Self Sufficient" therefore under your definition anyone under the age of say 12 is non "Self Aware" Self Awareness has to do with comprehending that the item/animal/person is what is doing/sensing/thinking something at that time. If the entity in question has the ability to understand and modify its action based on what it is receiving from its senses, Then the entity can act accordingly and modify its behavior and FUTURE behavior (however long or short that may be base on memory length) to either repeat or avoid the stimuli it received when it performed said action. Thanks, T
-
Your "Self Awareness" sounds to me more like "Self Sufficient" therefore under your definition anyone under the age of say 12 is non "Self Aware" Self Awareness has to do with comprehending that the item/animal/person is what is doing/sensing/thinking something at that time. If the entity in question has the ability to understand and modify its action based on what it is receiving from its senses, Then the entity can act accordingly and modify its behavior and FUTURE behavior (however long or short that may be base on memory length) to either repeat or avoid the stimuli it received when it performed said action. Thanks, T
Excellent point. Yes, self sufficient is more what I am thinking. I like your definition of self aware. A +5 for that :)
""Real programmers just throw a bunch of 1s and 0s at the computer to see what sticks" - Pete O'Hanlon
