Discussion in 'Intelligence & Machines' started by roadblock, Apr 6, 2006.
i got the link on the other post. well, since you asked:
Log in or Sign up to hide all adverts.
After looking at the site it seems you're just another guy who thinks YOU'VE found the solution (though you don't seem to have implemented it yet). In reality, you've published no papers on the subject, have no proof that any of your methods would result in anything resembling intelligence, and in FACT... it looks like you're just trying to make money from your books. This is not how science is done, and it is certainly not how you gain respect in the research community.
probably true, but I do indeed respect the fact that he's doing more than talking smack like myself. at least he's actually trying, writing code - pursuing the dream. i don't know if he can get to human level ai with what he's doing or not, but damn I really dig it when people try to solve what are supposed to be unsolvable problems.
"In reality, you've published no papers on the subject, have no proof"--antonk
are you kidding me? look at the pseudo codes/website pages, thats research. look at the three books, thats paper work.
and as far as i'm concerned this isn't about money. if this was about money i would of quit a long time ago. what makes you think someone would spend 7 years working on something without getting paid one cent? this is an interest for artificial intelligence and a genuine desire to see the development of human level AI robots.
as far as the proof, i got some on the AI demos. those are simple versions. the next wave of programs will be coming out shortly. you have to understand that the universal artificial intelligent program is very long to write. it can't be done overnight. it takes time. but given the time it will be completed and the function of the program will be the proof.
Webpages and pseudocodes do not constitute peer reviewed research publications. There are plenty of conferences and journals that would take real look at your work and decide if it was worth publication.
"There are plenty of conferences and journals that would take real look at your work and decide if it was worth publication."
what????? i assume that only professors at universities are the only ones who are smart enough to build human robots and regular people don't have that kind of knowledge.
no one has an idea how the brain works. no one has the creative mind to even come up with a theory. my work is one possible theory. you can look at my stuff and either think it won't work or it will work. thats what science is. not some 100%-proven-mathmatical-theory that is published in science-magazines.
point...but isn't consciousness defined in terms of understanding and not the other way around?
Isn't this the soul? I thought we didn't believe this Please Register or Log in to view the hidden image!
As for you 'meaning' dilemma, meaning can be expressed in terms of data and information. Links between mental objects allow definition and inheritance of properties; I think meaning/data/information/knowledge can be expressed in terms of system-states.
What say you?
Please Register or Log in to view the hidden image!
In the context of trying to understand understanding, no I don't think so. You could say that and be correct from one point of view, but from another... "a definition" really only exists in a mind as part of it's own context. Until it is read and integrated into the individual, it is simply meaningless unknown... completely undefined.
People call it lots of things. I personally find the notion of "supernatural" wholly beside the point of understanding nature. That which is not natural, is not possible. Thus, if it exists - it was possible and thus - wholly natural.
It can also be expressed as "an experience". Meaning can be conveyed to a large extent with data and information... but until data becomes information, it is only potential. It is of no meaning, whereas information is as in "someone has been informed". As such, data and information are two different things (to me at least) which describe a transition rather than a relationship that comprises meaning. Meaning is to me 'impact upon a psyche' in the sense that 'it is experienced by the observer as its internal reality'.
Certainly, but what exactly comprises the elements of the system? I think "the capacity for abstraction" (basically the meaning thing I've been after) in the sense that 'self' is projected upon a pattern 'observed' in its perception, or perhaps in the collision between what has been "observed' and what is being observed. IMO, there is something more to the present than just a nearly non-existent, infinitessimal hypothesis of a moment, slammed between the past and future. That infinitessimal slice is where "I" always am. Further, what is it exactly that finds or relates patterns?
There is only tao (the whole, unperturbed, unobserved). Something within that whole separates bits of it into "categories" that ultimately do not exist within the space-time aspect of "the whole". Certainly the object that was categorized likely "exists" as to be percieved... but the category, as it 'things can be put into separate groups with descriptions that speak to their attributes' is an entirely different issue - at least in my view.
Again, could be the thick skull.
"Isn't this the soul? I thought we didn't believe this"
the conscious is not a soul. it is simple the activation of objects (things) that have association with what we are sensing with our 5 senses. another thing is that the consious is also caused by patterns that the brain found.
there is no soul that guides the body and tells it what to do.
"I think meaning/data/information/knowledge can be expressed in terms of system-states."
not system states but more like patterns and rules. the 5 senses coming into the body is broken down into objects and stored in the brain based on what is familiar. for example, if i learned what a computer looks like then the next time when i see a computer i would think that the computer is one object.
it does follow the traditional way of AI which is that of system states. these states have their own rules and conflicts.
The planning part of it is also pretty much the same. current states, and goal states but the brain universalized everything and learn to plan a task on its own. on the other hand most planning programs require that the program be fed manually the rules and the conflicts. the human brain learns these rules and conflicts on its own.
for example, if i learned what a computer looks like then the next time when i see a computer i would think that the computer is one object.
but it's also "a computer", "a machine", 'blue, green and black", "a thing", "something that weighs about 20 lbs", "a device for expanding communicative abilities", "a processing unit", "something that is obscure to an ant", etc.
the thing about you approach that I think falls slightly short is well, maybe I don't quite understand it, but I think its that what is related in a mind is not just "information", it's "concepts" which are IMO, basically an abstracted chunk of direct "experience in time" inclusive of emotional contebt and which yeilds different information dependent upon the current context, or "momentum of thought".
can a program, for the simple joy of musing - hold up the idea of "a computer" and look at it from as many angles as it can think of, just because it felt like it at the time? can it "feel like it"?
can it doubt that it is askign the right questions, or that it really even understands what it encounters?
you are refering to the details of what a computer is. that is something i only touch on basically in my site. but to some up a computer, it is a visual object with 3-d characteristics. you can look at a computer from the front, back, side, angle, 100 meters away and it will still be considered a computer.
i don't wnat to go into the really really details of how this works.
the conscious is easy to sum up and easy to test if i'm right or not. the conscious is basically activations of other objects associated with what you are sensing. here are some examples. on the left are the sense objects and to the right of the equals is the activation objects.
words = meaning
bill clinton = sex scandal/monica luinsky
george bush = war on terrorism
bill gates = microsoft/richest man in the world
toyota = car company
The objects to the left are called target objects and the objects to the right are called the element objects. The target object will activate another object (element objects) in memory that have strong association. This varies from individual to individual but most of the information is the same.
the conscious not only does this but it tells us things that we have to do or things that will accomplish a task. when people are on the cross walk the first thing that pops up is: look left and look right before crossing the street. this "look left and look right before crossing the street" is one object that has strong associations with the cross walk. (we were taught this over and over again). this is why it gets activated when people want to cross the street.
meaning to words and sentences are also understand by repeatedly learning the meaning. some of it is learned by patterns the brain found. patterns is very important in terms of learning anything. This is why when we see a word or read a sentence the strongest assocation is the meaning. thats what gets activated in the mind.
Instead of 'the joy for musing', it'll have a positive feedback loop that increments it's happiness counter by an amount equal to its curiosity coefficient (a constant) along with other, randomized factors. It can also compare the properties of different objects ('Computer', 'cheese', 'I') and try and find similarities between them so it can predict properties of the afore-mentioned objects.
"abstracted chunk of direct "experience in time" inclusive of emotional contebt and which yeilds different information dependent upon the current context, or "momentum of thought"."
well, thats what the brain does it modifies information and forget and learn information. words like "bad" isn't fixed for eternity. it changes as time passes. when michael jackson came out with the phrase: "those dance moves are bad". he is actually changing the meaning of the word "bad". it went from a bad thing to a good thing. we learn these things through experiences and those experiences change our mind and what we percieve the world to be.
basically a society decides what things are. because thats where humans get all their information from. the way we dress, the way we speak, the way we interact with each other and the way we behave.
i think what kron is refering to is the desirability of objects. for example, a man would see an attractive women with high points. while a man will see an unattractive women with negative points. and these things also is factored into the way we make decisions.
What if there are degrees of sentience?
"and be correct from one point of view, but from another... "a definition" really only exists in a mind as part of it's own context. Until it is read and integrated into the individual, it is simply meaningless unknown... completely undefined."
actually meaning is something hidden. it is only patterns and can't be seen. something like a "computer" that has many meanings and many patterns associated with the word. you can close your eyes and hear the windows operating system opening music and think computer. you can look at a magazine with a pciture of a computer and think computer. you can see a cpu and think computer. The words "computer" makes you think computer. The brain takes the strongest meaning to computer and output that. only one can be outputed at one time (they compete to be activated). and it doesn't output the same meaning everytime. it really depends on what kind of senses that the human is sensing that causes the appropriate meaning of computer to activate.
"What if there are degrees of sentience?"
like i said there are many meaning to words and sentences. the word "site" have multiple meanings both directly and indirectly. site could be a website. site could be an environment. site could be eye site. site could be a picture etc. It is up to the senses that the person is witnessing "at the moment" that define what the meaning of site is.
Can a human being know? And why?
I'm sure if this is answered, then it'll make it easier for some. Of course, I said some, not everyone.
It depends on what you mean by know.
A human being can be aware.
And knowing is to be aware of "knowledge"?
What is knowledge then?
Is "to know" absolute?
In one sense it is, in that to "know" is to not "not know', so it's either "knowing" or "not knowing". In another sense, what is known can be questionable, unsure.
I think knowledge is basically, part of self and as such - humans can know. Can "self" be implemented in a machine. That's the hanging question to me. I don't think it can at this juncture, because the capacity to internalize an "awareness" relies on more than can be explained in classical space-time. I think you have to reach to m-theory type stuff to understand it. I've gone on long enough in other threads as to exactly why.
... but there is an implicit point in this. meaning only means something to an 'awareness'. A program implements code. It is not "aware" of the patterns. It executes commands on them.
I'm not arguing that your ideas are wrong in the sense of being able to provide good answers to specific 'profit functions', but that your program can't be aware if itself in a "self'-meaningful way in its function.
Then again, maybe you have a block diagram for the aritificial ego. That would be cool!
Can you write the profit function in such a way that it understand clearly that it is "itself"?
To me the concept of an inalienable and irreproducable 'consciousness' is equivalent to a soul. It is a delusion that we share to convince us that we are special in the universe.
Here's my idea in theroem form:
Postulate: All thought processes, human, machine, or otherwise can be explained in terms of physical phenomena.
Postulate: Any physical phenomenon can be replicated within the controlled conditions of a lab.
Conclusion: The human mind can be reproduced purely through mechanics. If we assemble a large group of molecules into an EXACT replica of a person's brain (along with the energy levels within said particles), we can produce an artificial mind.
Conclusion: Intelligence can be manufactured. 'Consciousness' too.
Well, you can call it that if you like. I personally don't like the term because religious people seem to be superstitious about it. My personal understanding of it is perhaps quite convoluted and irrelevant to other perspetives. Personally I feel that I'm insightful on the topic, but to someone else, maybe not so much.
The religious connotation, mostly yeah. It's not quite the reason it's done, but close enough.
Objection: There is no evidence to support this postulate, nor refute it. Thought processes are not completely understood. "physical phenomenon" are limited to the physical, having no "abstract component" and therefore cannot fully describe thought.
Objection: You can't birth a star in a lab, because the lab would be consumed in the process. Therefore your postulate is false.
Considering the objections, the conclusion does not follow.
I do think we'll grow to understand and "manufacture" it, but I think it'll be outside of the current paridigm of thinking about the problem, in that there is some aspect of the universe that still needs to be explored such that we might account for it in our experimental model of consciousness.
By definition, the "abstract" is not phsyical... yet "self" is both "real" and "abstract". I think it follows that "physical" does not allow the possibility of "abstract", and that "abstract" is utilizing a degree of freedom within the universe that is only hypothetical at this juncture.
Consider mackmack's point that "meaning" is "hidden" in patterns or whatever. Why? Where exactly is it?
Separate names with a comma.