Unrestricted system

[tkorrovi]

By unrestricted i mean a system where any system can emerge within a bigger system, as a result of self-development, generality by McCarthy or type 0 grammar by Chomsky.

I don't claim that Absolutely Dynamic System is unrestricted, as it is not proved that it is absolutely dynamic, and so also are not proved the other things which follow from that. However, the Absolutely Dynamic System was derived to be as dynamic as possible.

But the question is not about an exact proposed mechanism, it is about creating an unrestricted system. Some read about unrestricted system, Artificial Consciousness etc, and then continue to study some other system, which someone claimed to be some kind of right solution, though it is restricted. And they don't think more about the main problem, whether a true AI should be unrestricted or not. But this is imho the most basic question, from which we should proceed, and one goes the wrong way when trying to ignore it. And i never heared any good argument, why true AI shouldn't be unrestricted. Some just say that they don't believe that it's possible to create an unrestricted system, when it is so, then why not create a system which is as unrestricted as possible? The only argument which i know, why they claim that an universal system should not be unrestricted in its basic mechanism, is the argument that all what is necessary for an universal system, is complexity, this argument can be represented in many different ways, like, mind is merely a set of behaviors, brain is like a computer with many programs etc. But it should be clear that it is not possible to achieve whatever with whatever kind of complexity, so complexity alone is not the answer. And some don't understand what unrestricted AI system means, they don't understand that unrestricted means unrestricted every way, like, when the system has a pre-programmed goal then, however unrestricted it is otherwise, it is most likely restricted because of goal, like i know no pre-programmed goal, which would not restrict the system. Also many other things, like pre-determining the dimensions, restrict the system. So, it is even not so easy to comprehend, what unrestricted means, unrestricted has a kind of immense meaning.

So, whether the system should be unrestricted, is therefore the first philosophical problem, which one should be think about, before going any further in creating any true AI system.

[raffethefirst]

I will ask for a favor if you dont mind.

Could you try to explain what unrestricted is?
You said that is difficult to describe it.
But this is my request.
Please give it a try as I simply cannot follow your theory without understanding what unrestricted is.
This is the word you use in all you phrases and connect all ideas.
You do understand that all people reading this and not knowing what unrestricted is will not understand all those phrases and ideas.
I think you should elaborate on this one.

Thanks

[tkorrovi]

Unrestricted system is a system which is able to learn whatever behavior, without having any prior knowledge about that behavior. Which means, it may have some general knowledge, but not a particular knowledge about the behavior which it is going to learn.

Unfortunately, even McCarthy writes about generality, but doesn't define what generality is. One should read his paper about generality though www-formal.stanford.edu/jmc/generality/generality.html to have more idea about unrestrictedness. McCarthy analyzes several different systems, and makes the conclusion that they still lack generality. Like, he describes the system where learning is accomplished "by making random modifications to the program and testing the modified program", and explains why such system lacks generality. One should read such papers first, before trying to create a true AI system, but still many repeat the mistakes which were foreseen already 20 years ago.

Chomsky defines type 0 grammar, or unrestricted grammar. where are no restrictions on the production rules. So one should also consider, that a true AI system should be Turing complete. Our programming language is Turing complete, but often the systems which we create using that programming language, are not. ADS is derived to have a type 0 grammar, enabling any system to emerge from a bigger system, being at that self-developing, what not every Turing machine is. An unrestricted system should satisfy such conditions. Unfortunately often, people who claim that they created a true AI systems, like Arthur but also much more responsible people, when asked, what reason they have to consider that their system is Turing complete, either have no idea whatsoever why this is important, cannot answer anything, or then say that their system is not Turing complete because it is too difficult to create such, like Arthur said, or such. Cellular automaton is Turing complete, but it is not derived so that it was Turing complete by having a type 0 grammar, which means that it is not anyhow substantiated that whatever system can emerge in cellular automaton, as a result of development of some pattern, ie by means of deriving a pattern from some other, infinitely developing pattern. Which means, in spite that Turing machine can be implemented in a cellular automaton, it is not substantiated anyhow, that any kind of structure implementing Turing machine can emerge there, as a result of self-development. That knowingly no one succeeded to train cellular automaton, further confirms that weakness of cellular atomata.

Again, why i talk so much about cellular automata, as some say they are nothing bout some game, and not important. This is completely wrong, in spite that cellular automata were used like some games, they have a very big theoretical importance. They are the only widely known universal self-developing systems. ADS is another kind of such system, and ithere are likely no others. I think it is also evident, that an unrestricted system should be self-developing in its basic mechanism.

[tkorrovi]

Well to say it more simply when this was too complicated... Say that you have a box, with many wires coming out of it. Now when you can train this box to have whatever behavior, by sending pulses into these wires and observing the pulses which come out of these wires, then this is an unrestricted system. When you cannot train it whatever behavior, but only a certain restricted range of behaviors, then this system is less unrestricted. But it should be as unrestricted as possible. The behavior is about sending and receiving pulses. This is, btw, how the brain works. only pulses go into or come out of the neurons, and as you can see, your brain can do great things, that way.

[tkorrovi]

And by training i mean, training just by interacting with it, no special purpose signals whatsoever.

[Trav's Buddy :)]

Just some examples of unrestricted system



www.x-art.com
www.celebritystorylibrary.com/

+

all inet so called ethical places

+

cyber crimes

+

propoganda about its severity and control checks 

+

anything you can think of...



They kinda complimentary for each other and more fortunately/unfortunately (which way u c it) now they obey this literal law .... and no sane or insane logic is suppose to pull this down. So they say we fight till death. haha! 

And yeah! they did and still do evolve from bigger system and there's lot of 0 type grammar goin on there as well, Whether chom type or otherwise...  not hard to guess I suppose   smile