|
Author
|
Topic: The Characterization of Intelligent Causation
|
Melvin H. Fox
Member
Member # 1684
|
posted 05. April 2007 22:46
aiguy,
I am perfectly content equating intelligence with intentionally seeking out, collecting, storing, organizing, synthesizing, analyzing, characterizing, and predicting information with respect to reality. And yes, equations are perfectly circular, a = b implies b = a.
I believe you have at least implied that your pet AI systems can learn. What are they learning about? Is it not what works (is right) and what does not work (is wrong)? Now if you have created for them some virtual reality then we should expect their characterizations to converge on that reality. This is true because if these systems can learn then they are self correcting. They are making corrections against the virtual reality. Therefore, the more trials they run, then the more corrections they make. And, the more corrections they make, then the closer to the target (given virtual reality) they should come. This is, of course, unless the corrections are random (no learning). In that case, as is the case with the random walk experiment, we would expect them to diverge from the target.
I freely admit that I can’t demonstrate free-will. If you require me to do so before we go on in this discussion, then I respectfully withdraw.
I have had the experience of witnessing the delivery of my youngest daughter, six weeks premature (via c-section) because, as the doctor later discovered, there was a double figure eight knot in the umbilical cord. You did not have that experience. What fixed law can be demonstrated to specifically contribute to this difference in personal experience? I have an innate aptitude for mathematics? What fixed law can be demonstrated to specifically contribute to my possessing this innate aptitude? If you can demonstrate no such contributions I request we speak no more of fixed laws causing inconsistencies in human behavior.
If I can’t talk about something being “purposefully specified by some intelligence” then you can’t talk about something being “caused by chance.” Fair enough?
quote:
And yes, I am perfectly satisfied with the answer "we don't really know" what causes biological complexity. However, nobody attempts to explain anything with chance alone.
There are plenty of theoretical physicists who explain the boundary conditions for our universe and the origins of all fixed laws in the universe by chance alone. Many physicists also explain the most fundamental quantum workings of our universe by chance alone.
quote:
And nobody should attempt to explain anything by "intelligence".
Your closing statement reveals your a priori assertions with respect to intelligent design. I have my own set of a priori assertions – very different from yours. I can’t see that we will ever agree on much.
-Mel
IP: Logged
|
|
Melvin H. Fox
Member
Member # 1684
|
posted 05. April 2007 22:55
Stephen,
As much as I respect 2ndclass, I doubt he will acknowledge the validity of the phrase “purposeful adjustments to a preconceived target state” even though he himself has already made purposeful contributions in this discussion and brings his own preconceived conclusions to the table.
-Mel
IP: Logged
|
|
aiguy
Member
Member # 3736
|
posted 06. April 2007 04:08
Daniel,
quote:
That's actually the whole point of the filter - to show that both watches, and protein folding, are most likely the product of design (I'll leave off the word "intelligent" since you have problems with it).
I think you missed the point here. Proteins fold up inside of cells, and they find the correct conformation out of countless possibilities in a short time. Since we can't explain how this happens by any known combination of fixed law and chance, the filter says that the result of this process - the folded protein itself - is the product of intelligence. Therefore the thing inside the cell that is folding this protein must be intelligent!
Now, do you think that the thing inside the cell that makes proteins fold up right is the same thing that enables us to design watches?
IP: Logged
|
|
aiguy
Member
Member # 3736
|
posted 06. April 2007 04:31
Melvin,
quote:
I am perfectly content equating intelligence with intentionally seeking out, collecting, storing,
organizing, synthesizing, analyzing, characterizing, and predicting information with respect to reality.
And yes, equations are perfectly circular, a = b implies b = a.
I'm afraid you missed the point.
1) Humans can do these things because they are intelligent.
2) Humans are intelligent because they can do these things.
We are not saying these statements are equivalent. The first attempts to explain why humans can do these things in terms of intelligence. The second explains why we call something intelligent - which is merely because they can do these things!
quote:
I believe you have at least implied that your pet AI systems can learn. What are they learning about?
Machine learning is a well-developed discipline - they learn about all sorts of things (Google "machine learning").
quote:
This is, of course, unless the corrections are random (no learning). In that case, as is the case with the random walk experiment, we would expect them to diverge from the target.
It's not a random walk, and the "corrections" (inferences) improve the system's ability over time, yes.
quote:
I freely admit that I can’t demonstrate free-will. If you require me to do so before we go on in this discussion, then I respectfully withdraw.
In that case, we have no way to demonstrate that anything acts by any other cause but fixed law and chance. Much of ID literature is based upon the assumption that we can so demonstrate.
quote:
What fixed law can be demonstrated to specifically contribute to my possessing this innate aptitude? If you can demonstrate no such contributions I request we speak no more of fixed laws causing inconsistencies in human behavior.
If fixed law is all there is, then fixed law accounts for everything. We have discerned two types of causes in everything science has studied: fixed laws, and things that happen with no apparent pattern at all (chance). ID's concept of intelligence requires a third type of cause which affects the material world; however, these is no evidence for this type of causation.
quote:
If I can’t talk about something being “purposefully specified by some intelligence” then you can’t talk about something being “caused by chance.” Fair enough?
Sure - I'll say "happen with no discernible pattern".
quote:
AIGUY: And nobody should attempt to explain anything by "intelligence".
MEL: Your closing statement reveals your a priori assertions with respect to intelligent design. I have my own set of a priori assertions – very different from yours. I can’t see that we will ever agree on much.
This is not an a priori assertion. I said this pursuant to the point I made at the top of the post. If we say only intelligence is that which can do things that we call intelligent, then we can't explain anything on the basis of intelligence. In order to give the explanation scientific content, you must characterize intelligence such that it can be objectively detected.
So no, I'm making an argument, not an a priori assertion. If you have a set of a priori assertions, however, then you're probably right - we aren't going to get very far.
IP: Logged
|
|
LifeEngineer
Member
Member # 3446
|
posted 06. April 2007 10:06
AIGUY: And nobody should attempt to explain anything by "intelligence".
MEL: Your closing statement reveals your a priori assertions with respect to intelligent design. I have my own set of a priori assertions – very different from yours. I can’t see that we will ever agree on much.
--------------------------------------------------------------------------------
aiguy: This is not an a priori assertion. I said this pursuant to the point I made at the top of the post. If we say only intelligence is that which can do things that we call intelligent, then we can't explain anything on the basis of intelligence. In order to give the explanation scientific content, you must characterize intelligence such that it can be objectively detected.
This exchange lends support to my earlier comments that it is not possible to resolve the scientific acceptability of a definition of intelligence unless you first resolve 1)the framework in which the definitions are to be presented, 2)the formal criteria involved in evaluating the definition, and 3)the set of evaluating involved in accepting or rejecting proposed definitions.
A lot of scientists involved in the study of intelligence might be willing to accept that the concept of intelligent causation is in some way associated with the concept or behavior of problem solving. Most scientists would, I believe, recognize that 'problem solving' is not a sufficiently precise definition to be used in scientific analysis. Lots of scientists and non-scientists can come up with lists of properties they believe should or should not be included in a more precise definition. But no amount of discussion will ever result in an agreement on which properties should or should not be included. Aiguy and people who think like he does can always reject any proposed definition as inadequate and people with different beliefs can just as firmly believe they have an acceptable definition.
The only way to resolve the issue of defining intelligence from a formal hard science perspective is to go back and develop formal rigorous mathematical/logical frameworks for expressing a definition, then determine if the defintion developed leads to 'best practices' predictive theories and best practices solutions to practical problems.
Discussing definitions of intelligence without getting into formal mathematical concepts and formal predictive theories, might be fun and it might mean that a lot of people without a working knowledge of formal mathematical and formal scientific methodologies can participate, but ultimately such informal discussion can never lead to resolution or agreement.
IP: Logged
|
|
LifeEngineer
Member
Member # 3446
|
posted 06. April 2007 10:33
To briefly review, I am proposing that intelligent causation be analyzed in terms of or using the framework of mathematical models of the general type- “F(S)=R where R is either goal compatible (GC) or not goal compatible (NGC)”
As aiguy has pointed out, the F(S)=R component of this mathematical structure might be labeled a Turing machine or might be characterized as Turing computable. Adding the GC/NGC classification or property of R to the model is something of a modification to the formal or abstract Turing machine concept, but it is a modification that is modification compatible with practical applications involving computer programs (people writing programs routinely evaluate whether the output of a program is or is not compatible with the goal of the program).
Given this abstract framework, it is possible to 1) develop a formal definitions of both intelligent and non-intelligent behaviors based on ‘efficient transformation of processing to produce R’s that are GC’ and 2) it is possible to model a wide range of real world behaviors and determine if the behaviors qualify as intelligent or non-intelligent.
As should be obvious to anyone who understands the mathematics and the modeling involved, the definition involved will end up classifying most relevant human behavior as intelligent, most evolutionary change behavior or processes as intelligent and most computer behavior as intelligent. Is there anyone who disagrees?
The next question is whether this formal definition of intelligent causation is useful in formulating hard science predictive theories. The answer to this question depends on an understanding and recognition of teleological theories or intelligence based teleological theories. Any questions?
IP: Logged
|
|
aiguy
Member
Member # 3736
|
posted 06. April 2007 12:43
LE,
I think we agree that in order for ID to be a meaningful idea, the notion of intelligence has to pinned down beyond our intuitive notions about free will, consciousness, and that the concepts that psychologists use to talk about human (or other animal) intelligence can't be applied in the context of ID.
But I'm confused about your proposal...
quote:
As aiguy has pointed out, the F(S)=R component of this mathematical structure might be labeled a Turing machine or might be characterized as Turing computable. Adding the GC/NGC classification or property of R to the model is something of a modification to the formal or abstract Turing machine concept, but it is a modification that is modification compatible with practical applications involving computer programs (people writing programs routinely evaluate whether the output of a program is or is not compatible with the goal of the program).
No, it is not that one or another aspect of your model is Turing computable, since every formal model is Turing computable. You have given some simple formulas here, but formulas do not represent "modifications" to Turing computability. "Turing computable" means "computable by some possible computer".
quote:
The next question is whether this formal definition of intelligent causation is useful in formulating hard science predictive theories. The answer to this question depends on an understanding and recognition of teleological theories or intelligence based teleological theories. Any questions?
The question is whether or not there exist any formal models of human behavior, and of the origin of biological complexity. Unless these models exist, then we don't know what we're talking about here. It's fine to say that formal models would satisfy the question of the OP (they would), but what are these models?
There is, of course, one proposal for a model that is supposed to explain both speciation and intelligence: Random generation plus selection by the environment. "Neural Darwinism" has some prominent proponents (Gerald Edleman, William Calvin, etc) - these folks believe that Darwinian processes in the brain are what enable us to solve problems, and of course evolutionists believe that analogous processes cause biological complexity. As it happens, I don't believe that either of the conjectures are true, and I'm certain that this is not what ID proponents have in mind. However, genetic algorithms can solve problems, and they can model biological evolution, so this would be one candidate.
You are offering another candidate, where you represent goal-oriented behavior. If you can show some of these models solving problems, and plausibly modelling the creation of biological complexity, then you've got something. You make a strong claim, viz. "It is possible to model a wide range of real world behaviors and determine if the behaviors qualify as intelligent or non-intelligent." That would be great - but until you actually show these models, it's just speculation. AI researchers strive to model real world intelligent behaviors, with some success, but no common model at all.
[ 06. April 2007, 12:44: Message edited by: aiguy ]
IP: Logged
|
|
2ndclass
Member
Member # 1979
|
posted 06. April 2007 15:03
Daniel: quote:
The question is a good one. How would we go about determining if pyramids (or any other structures) on Mars were designed?
Ah, but that isn't the question I was asking. I was asking how we define the set "intelligent designs", not how we determine whether things are designed. Determinations can be right or wrong; definitions are neither right nor wrong. If we can delineate a set of "intelligent designs" and determine that they all came from the same cause, then we can define "intelligence" as that cause.
Alternately, we could define "intelligence" first, and then define "intelligent designs" as whatever is produced by intelligence. But the ID movement doesn't seem to be taking either route.
One of the main points of this thread is that the ID movement is putting the cart before the horse. They are proposing methods for detecting intelligence without bothering to define the term first. The word carries subjective connotations, but that's not good enough for constructing a scientific hypothesis. ID has yet to rise above the level of nebulous philosophical talk.
Melvin: quote:
I thought long and hard about the above comment. No, 2ndclass, I don’t want to go down that route. I have free-will but not unlimited power. If that is the route that human behavior is headed down, I can’t stop it. I can try to explain it. My explanation is free-will.
The "route" I was referring to is acceptance of the idea that reality can be logically incoherent. I still can't figure out where you stand on this.
Stephen: quote:
Intelligence is indicated – strongly – by phenomena that exhibit use of feedback to guide events toward a goal or pre-imaged structure.
What do "goal" and "pre-imaged" mean in a non-human context? Can a rock rolling down a hill be said to have a goal of reaching the bottom? (Note that such a system has feedback -- if the rock starts rolling in a direction other than straight down, gravity corrects its course.) Is a computer program an image of its future output?
And, more to the point of this thread, how can we say that something indicates intelligence if we don't know what intelligence means?
Melvin: quote:
As much as I respect 2ndclass, I doubt he will acknowledge the validity of the phrase “purposeful adjustments to a preconceived target state” even though he himself has already made purposeful contributions in this discussion and brings his own preconceived conclusions to the table.
I have no problems with the words purposeful, preconceived, or target when we're talking about humans. I just don't know what they mean outside of that context.
LE: quote:
Any questions?
Yes, two of them:
1) What are your definitions and model?
2) Why bother asking if there are any questions when you ignore almost every question that's put to you?
[ 06. April 2007, 17:06: Message edited by: 2ndclass ]
IP: Logged
|
|
Melvin H. Fox
Member
Member # 1684
|
posted 06. April 2007 15:07
quote:
1) Humans can do these things because they are intelligent.
2) Humans are intelligent because they can do these things.
We are not saying these statements are equivalent. The first attempts to explain why humans can do these things in terms of intelligence. The second explains why we call something intelligent - which is merely because they can do these things!
NO! When I say “Humans CAN do theses things because they are intelligent”, I am NOT attempting to explain why humans do anything. That is a different question. I am trying to equate these activities with intelligence and establish the fact that humans do these things inconsistently. Perhaps I could have said it in a better way but that is what I was trying to say. If we want to find out the cause of human behavior, then we will have to determine why their behavior is inconsistent. If the inconsistencies are due to chance then these inconsistencies should filter out via trial and error. If they do not, then either humans can’t learn or there is some other explanation. To ignore purposed free-will as a possible cause is revealing of a priori assertions.
-Mel
IP: Logged
|
|
nosivad
Member
Member # 767
|
posted 06. April 2007 15:17
We are all victims of our "prescribed" fates.
"EVERYTHING is determined...by forces over which we have no control."
Albert Einstein, my emphasis.
Long before Einstein -
"Every boy and every girl,
That is born into the world alive,
Is either a little liberal,
Or else a little conservative."
Gilbert and Sullivan, Iolanthe, 1883
Hence the Prescribed Evolutionary Hypothesis.
The whole thing was planned from beginning to end and the end is now.
"A past evolution is undeniable, a present evolution undemonstrable."
John A. Davison
"La commedia e finita."
Pagliacci, last words.
[ 06. April 2007, 15:36: Message edited by: nosivad ]
IP: Logged
|
|
aiguy
Member
Member # 3736
|
posted 06. April 2007 16:42
Mel,
quote:
NO! When I say “Humans CAN do theses things because they are intelligent”, I am NOT attempting to explain why humans do anything. That is a different question.
I understand you are not trying to explain particular instances of behavior. You are, however, trying to explain the fact that we can do various things (design a watch, write a book, and so on) by appealing to "intelligence". It is this that I am saying is vacuous.
Q: What enables humans to design watches?
A: Intelligence.
Q: What enables humans to pole vault?
A: Athleticism.
Q: What enables opium to put people to sleep?
A: The sleep-inducing properties of opium.
These explanations are all vacuous, because the explanation simply reifies that which we are trying to explain.
quote:
I am trying to equate these activities with intelligence and establish the fact that humans do these things inconsistently. Perhaps I could have said it in a better way but that is what I was trying to say. If we want to find out the cause of human behavior, then we will have to determine why their behavior is inconsistent.
Why do you think inconsistency is the defining characteristic of intelligent things? If I consistently gave the correct answer to every possible question, I would think that would be the epitome of intelligence.
quote:
If the inconsistencies are due to chance then these inconsistencies should filter out via trial and error. If they do not, then either humans can’t learn or there is some other explanation.
Let me see if I can make my point more clear here. Let's say I have a program that responds to questions like this:
Q: How are you today?
A: I'm good, thanks.
Q: How are you today?
A: I've been better.
Q: How are you today?
A: The tuna is a good fish.
Here, my program is acting inconsistently: Each time it is asked the very same question, it seems to give a different answer. How could a program do this? There are two basic ways:
1) The program picks an answer at random
2) The program remembers every question it has been asked, and is programmed to give a different answer when asked the same thing again.
The first is what I mean by chance. The second is what I mean by saying that our actions are determined by the totality of our prior experience. Some combination of these two are, as far as we can tell, what causes our behavior.
quote:
To ignore purposed free-will as a possible cause is revealing of a priori assertions.
The only a priori assertion I am making is that in order to make a scientific claim, we must be able to demonstrate its truth. Since we cannot demonstrate that anything transcends fixed law and chance, we cannot base our scientific theories on the assumption that contra-causal free will is responsible.
[ 06. April 2007, 16:45: Message edited by: aiguy ]
IP: Logged
|
|
aiguy
Member
Member # 3736
|
posted 06. April 2007 16:44
Hello there, nosivad -
Wow, I've never actually talked to a Leibnizian determinist before. Doesn't quantum uncertainly wreak a bit of havoc with these "front-loading" schemes?
IP: Logged
|
|
LifeEngineer
Member
Member # 3446
|
posted 06. April 2007 16:48
Quote: No, it is not that one or another aspect of your model is Turing computable, since every formal model is Turing computable. You have given some simple formulas here, but formulas do not represent "modifications" to Turing computability. "Turing computable" means "computable by some possible computer".
I have defined or identified a mathematical framework or perspective for modeling information processing behavior. That is essentially what Turing did and in fact the F(S)=R is the same framework defined by Turing. I am simply adding a classification of R or the GC-NGC property to the Turings abstract framework.
It should not be particularly difficult to understand 1) a general framework for modeling- as you agree the turing framework applies to all computer models, 2) a definition compatible with the defined framework that 3) makes it possible to classify modeled behavior as intelligent or non-intelligent.
I am proposing that all behavior be modeled by a slightly modified Turing compatible model. You seem to agree that all behavior modeled by computers involves a Turing type model.
Once you have a model of a particular behavior, you then determine whether the behavior is or is not intelligent based on the properties of the model.
There is nothing particular unusual or complex about this approach.
Quote: The question is whether or not there exist any formal models of human behavior, and of the origin of biological complexity. Unless these models exist, then we don't know what we're talking about here. It's fine to say that formal models would satisfy the question of the OP (they would), but what are these models?
I wonder how you reconcil this rather stange position with the obvious fact that there are all sorts of computer models in existance of all sorts of behaviors. It is quite contradictory to both claim there are lots of computer models and then suggest there are no such models.
The relevant point here is that we can start with any existing models of either human behavior and existing models of evolutionary change and if there is sufficient information available in the model we can determine if the model satisfies the proposed definition of intelligent behavior or the definition of unintelligent behavior.
IP: Logged
|
|
aiguy
Member
Member # 3736
|
posted 06. April 2007 16:50
Hi 2ndclass -
quote:
I have no problems with the words purposeful, preconceived, or target when we're talking about humans. I just don't know what they mean outside of that context.
I actually think that some of these words can be given clear meaning. Cybernetics defines "purposeful", for example, as a system that incorporates negative feedback. As we both know, however, this isn't very helpful in distinguishing things we intuitively want to call intelligent from things that aren't (evolution is purposeful in this technical sense).
IP: Logged
|
|
LifeEngineer
Member
Member # 3446
|
posted 06. April 2007 16:53
Quote: There is, of course, one proposal for a model that is supposed to explain both speciation and intelligence: Random generation plus selection by the environment. "Neural Darwinism" has some prominent proponents (Gerald Edleman, William Calvin, etc) - these folks believe that Darwinian processes in the brain are what enable us to solve problems, and of course evolutionists believe that analogous processes cause biological complexity. As it happens, I don't believe that either of the conjectures are true, and I'm certain that this is not what ID proponents have in mind. However, genetic algorithms can solve problems, and they can model biological evolution, so this would be one candidate.
But, of course, genetic algorithms involve intelligent or efficient searches rather than so called Darwinian chance variation and selection searches.
The random search or true random search criteria or standard involved in neo-Darwinian 'theories' is the same criteria I use for differentiating between intelligent and non-intelligent behavior. True random search is 'in theory' the basis for genetic algorithms, but in reality, genetic algorithms utilize intelligent or efficient search processes. Note that the most intelligent or most efficient search process is a process that finds the solution (GC form of R) on the first trial.
[ 06. April 2007, 17:03: Message edited by: LifeEngineer ]
IP: Logged
|
|
|
Printer-friendly view of this topic
