“We don’t actually know that other people are conscious.”
says The Science Channel, Morgan Freeman, host of: Through the Wormhole

Reference:
http://www.youtube.com/watch?feature=player_embedded&v=iDu5X-NMghg#at=1890

“We’ll have to rely on conversation.
If we ask the artificial brain, “Are you conscious?” and it persuades us that it is…
We’ll just have to take its word, but the same thing applies when we talk to other
humans. We don’t actually know that other people are conscious. It might just be
zombies who are saying the right thing, but having no private subjective experience.”

Professor Steve Potter of the Georgia Institute of Technology said, “I can put some
sort of an artificial intelligence into a computer right now that represents the kind
of decisions I might make and has in some sense some of my conscience in it.
So that’s not hard to do…”

Reference:
http://www.youtube.com/watch?feature=player_embedded&v=iDu5X-NMghg#at=2145

Hans Peter Willems - Jul 20, 2012:

Victor Shulist - Jul 20, 2012:

but we need examples Hans, examples.

You keep using the fact that I do not yet show my results in public, as some ‘proof’ that there are (probably) no results.  That still is a bit stupid , I don’t have any obligation to you to show my system, as to earn some rights to talk about it here.

  I guess that’s why you suggested it lol ... kidding. .come on Hans, calmn down, let’s communicate like adults shall we?  As in ‘we need’, I pretty much meant ‘we’d reallly like’ to see the results, because it sounds interesting.  wow, touchy one aren’t we.

Hans Peter Willems - Jul 20, 2012:

Victor Shulist - Jul 20, 2012:

Let’s get one thing straight… giving a computer rules *IS INFORMATION*... we give children information when we educate them.  There is no difference.

There is a BIG difference in that humans have the capability to discard the rules when they see a necessity to do so. Did you see the movie I Robot; the difference between Sonny and the other robots is exactly about that.

Entering information in a system either in raw data format, or data in the form if production knowledge (if/then), is does not preclude designing a system that can have that functionality.  Whether a specific type of information is entered into a well designed AGI system, be it, a series of integers, voltage or air pressure readings, or “if a noun is followed by a verb, it could represent a sentence”....  in a well designed AGI,  that information is just that INFORMATION .. . not *CODE* that the system is *made of*.  (That is, what I call “control code” is the code that *is* the system, and “considered code” which the system can look at and evaluate).

The control code tells the system how to deal with information in the form of rules given in natural language.. which it can validate against other statements to find for example, contradictions, or correlate with other information to produce its own complex natural langauge statements.

for a human, “control logic” would be the laws of physics that govern the chemical reactions going on during synapse firing.  That’s pertty low level, so I don’t think anyone can change their own control logic BY THINKING anyway. .. . yes, when we figure out the entire brain, we can do surgery to correct for things, but you get the idea, well, you probably don’t but oh well.

“CONSIDERED” logic is not logic that dictates how the system *works*....  that’s why I call it considered logic… the system can consider it, and evaulate it in the context of what it already knows.

If user inputs statement, and later inputs opposite statement, the system should say, woow woow hold on here….  and ask for clarification , and update itself.  Very simple example,  RULE :If X is a bird, then X can fly”.

System takes that in, but later is given “A Penguin is a bird”  and also “Penguins cannot fly”.

The system should update its “rule” .. .that is update its CONSIDERED LOGIC

FROM —If X is a bird, then X can fly”.
TO—If X is a bird, then X can fly, unless X is a penguin.

By having these completely different realms, levels, types whatever you want to call them, of rules, you see that providing rules (considered rules) initially into that system, does NOT PRECLUDE the “I robot” functionality.

And yes, the next logical step is to have the generated/updated considered logic produce a new version of the system’s own CONTROL logic.

But either way, you must start with some control logic.  You can’t start with ZERO RULES (that is, zero control logic), since the CPU in the machine would do nothing at all. . .and your raw data would just in ram. . not much would happen. . . so, it is ridiculous to suggest that putting rules in a system precludes irobot functionality.

Now, a system that has ONLY control-logic.. that is every ‘rule’ is in the form of control logic and their is ZERO considered logic, and ZERO ability of the system to “do the i-robot thing” with any considered logic, then you will never have AGI.. only narraw A.I.

Out of time, I will come back with responses to your other stuff later.

Victor, I’d like to see you reply to the other two messages I posted.

=================================================
    IF-THEN rules and pounding a square into a wheel
=================================================

First off , read my post.  if those if-then rules you are refering to are CONTROL LOGIC, then yes,  you may have a point.  If you are talking about CONSiDERED logic, then no, you don’t.  Rules, logic, code, instructions, data, all these things exist inside a computer system or AI/AGI system at many levels. 

Level 0 - hard coded algorithms in the CPU itself—the logic circuirty of the physical transistors configurations.

Level 1 - microcode . . .pattern of transistor off/on patterns it takes to execute one CPU opcode.

Level 2 - the complete set of opcodes/assembly language/machine language instructions that constitutes a compiled program.

Level 3 - scripting code , Perl, Phython code that feeds into an already compiled program.

Level 4 - Information ... in the form of anything from sensory input raw data, to the Pythagorean theorem, to grammar structures, etc.

One could argue that humans don’t think.  Did we “write our own code”  .. that is the CONTROL code in our brains (logic, or in N.I. (natural intelilgence), the laws of physics of the chemical reactions in our nuerons) . .No ..we didn’t , so one could argue that we humans were GIVEN the ‘rules” (control logic, laws of neuron physics/chemistry).. thus we are not thinking.  Nobody thinks like this of course, we *allow* humans to at least start with the CONTROL logic of how our neurons work. . in order to learn CONSIDERED logic.

Question for you—- since you are against the idea of just giving the computer grammar information in order for it to understand language, answer me this please….

As a child, in whatever grade, it was so long ago now, I was taught how to add, by being provided the information in natural language.

So I wasn’t given, you know, huge piles of raw data examples of additions like 1+1=2, 10+10=20, 20+20=40 etc and infered how to add… they gave me an example, and walked me through it, with language… showed me how the operation worked.

Children are taught using language.    Now yes, in the case of children they don’t learn language WITH a language, because they start with no language.  But the computer is different.. it comes built in with many levels (see above) of languages, from very low level to high level.

Thus, it is not that i’m against learning language by raw data, but it isn’t necessary, the computer has the power and built in abilities from the start that we can use a more direct, and more effective approach.

Even with the way I’m going about taclking language, I am discovering just how deep and complex it is, I’m not sure if you are aware of it.  NLU has been one of the jewels of computer science now for over half a century . .it is NOT an easy problem and I guess I really doubt if probabsllistic methods are going to cut it.  I know there HAS been some success with it, and I’m familiar with the techniques.  But here we are, this long after and we don’t have a system that can pass a turing test… is probalistic methods alone good enough for the task?

So Statistical Machine learning, although kind of cool, and yes, has proven effective in a few cases, google’s driving car using ANN’s for example, is pretty nice… but that is nto the way children learn.

And yes, it can learn to PREDICT. .but I really dbout that simply predicting things is the same as CONCEPT LEARNING.  I can’t see it.

Thinking about this a bit more, when I said children don’t learn language from language is not 100% true, we do have a ‘body language’ and many other forms of non-verbal communication.  So perhaps children DO learn natural language from a lower, more simple form of langauage.  And this is my approach, I have developed a language for teaching the system higher level languages like English . . but providing it the information it requires to make sense out of the given language.    Humans I think do have some ‘built in’ language.  For example, without knowing any English if a child does something inappropriate the parent may shout BAD !! !! BAD !!  . .the word means nothing but it is the fact that it is shouted, the tone of voice. . .the child knows innately that a harse tone is a negative thing.  HOW ?  this is hard coded. .this is a human’s equivallent to Level-0 or Level-1 language in a computer system, assembly language opcodes.

Now . .. why not take that a step further and develop a more powerful effect langauge which we can explain to the computer , a direct, effective way to explain higher language like English in a more simple language? 

children DO learn by examples yes, and there probably is some things in common with machine learning algorithms, but there is another component, innate knowledge that children have, that is hard coded.

I see starting with a boot strap language like English, and allow the system to learn other concepts from there, learn new CONSIDERED logic from there.  learn toher NLs from there, and perhaps learn to program via considered logic.. .  .and the ultimate goal is to have it use, as it pleases (the “I robot” functionality) this considered logic to produce its own CONTROL logic.  Write other programs, or rewirte sections of itself.

stupid chatbots.org timeout on edit .. .  harse >> harsh above lol.

Victor, you can go on and on repeating your views (which are still pretty narrow to my opinion), it doesn’t change the fact that the roundness of a wheel goes to the utility of the design, in the same way that programming a computer goes to the utility of the design of a computer-program and therefore the utility of the computer running that program. You simply fail to see the ‘roundness’ as programming of the wheel.

Let’s move a small step from wheels towards something that looks more like programming: clockworks. The pegged wheels, springs and cogs inside a clock are perceived by many people (including scientist) as early forms of programming; so again by your analogy clocks would ‘learn’ how to count the time by means of instantiating the functionality by the clockmaker (clock programmer).

It really doesn’t matter if learning is done by language, experience or whatever other means. To be able to learn, you need the ability to communicate, either in language or with other means in the absence of language (e.g. deaf and/or mute people). The form of communication used in learning does not ‘define learning’. The way that a system handles the information by means of comprehension DOES. A system learns when it can process information in such a way that it can leverage it’s utility BY ITSELF, based on the processed information. In software, this means you either need a system to be able rewrite it’s own software (which is a messy solution and very hard to control), or you need a system that can form new conceptual information BY ITSELF, based on the available conceptual information, that way creating richer conceptual models of reality.

However, you seem to avoid my request to reply to the other things I posted.

“. A system learns when it can process information in such a way that it can leverage it’s utility BY ITSELF, based on the processed information. “.

yes, that is what i was trying to get across.  We actually seem to agree on WHAT it should do, perhaps just differ on HOW.

For the relational understanding versus ‘in reality’ . . I won’t waste my timie.  That is a very philosophical topic with NO PROOF of any ideas we both may put forth . .and I know you ..you’ll demand you are correct on these deep issues that no one has answers or proof for   Don’t have the time for that.  Carry on.

Another thing we disagree on Hans, is you care about what the code or design *IS*.  I care only what it DOES.  To me if someone wants to call some code thinking or not is absolutely insignificant.  Some argue that a calculator doesn’t really calculate !!  And you know what, when thought about a certain way, they are perhaps correct, but who cares.. it does the job better than a human.  To me , that is the bottom line.  If you want to consider your specific design as ‘thinking’ and no one elses, that is fine with me sir.    It makes no difference what it is , or what consider it to be.  It is what it DOES that will matter.  Intelilgent people (and investors, although a lot of investors are far from intelligent), would or SHOULD judge a product by what it does, not on ridiciolus philosophy.

Victor,

If you want to consider your specific design as ‘thinking’ and no one elses, that is fine with me sir.

Hmmm… I don’t recall me ever saying that. I never stated that ‘no one else’s design is thinking’, just that purely rule-based systems (like e.g. a clockwork, calculators and indeed expert-systems) are not thinking. Mind you, I’m not in the scientific minority with my view on this, you are. Your replies to me are suggesting that I’m holding some awkward view, while my views on this are pretty much in line with most established scientists in the fields of cognitive science and AI-philosophy. Yours clearly are not.

Yet you still evade my request as to the information from Ben Goertzel and Margaret Boden (who are of course both such established scientists).

“The pegged wheels, springs and cogs inside a clock are perceived by many people (including scientist) as early forms of programming”, said Hans, and to that I admit, “I am one of those many people.” referring to the Antikythera Mechanism, the oldest computer, around before Jesus Christ historically, politely speaking.

Please go on about analogue computer programming… It’s a very interesting and useful for studying ancient computer programming, fundamentally unchanged for millennia.

What is thinking? (definitions from the web)

 Mind
    Thought (choose one or more)
        Cognition
        Consciousness
        Imagination
        Intelligence
        Perception
        Reasoning
        Others?

Mind is the complex of cognitive faculties that enables consciousness, thinking, reasoning, perception, and judgement — a characteristic of human beings, but which also may apply to other life forms.

Thought generally refers to any mental or intellectual activity involving an individual’s subjective consciousness. It can refer either to the act of thinking or the resulting ideas or arrangements of ideas.

Cognition is a group of mental processes that includes attention, memory, producing and understanding language, solving problems, and making decisions. Cognition is studied in various disciplines such as psychology, philosophy, linguistics, science and computer science. Usage of the term varies in different disciplines; for example in psychology and cognitive science, it usually refers to an information processing view of an individual’s psychological functions. Cognition is a faculty for the processing of information, applying knowledge, and changing preferences. Cognition, or cognitive processes, can be natural or artificial, conscious or unconscious.

Consciousness is the quality or state of being aware of an external object or something within oneself.  It has been defined as: subjectivity, awareness, the ability to experience or to feel, wakefulness, having a sense of selfhood, and the executive control system of the mind.

Imagination is the ability of forming images and sensations when they are not perceived through sight, hearing, or other senses. Imagination helps provide meaning to experience and understanding to knowledge; it is a fundamental faculty through which people make sense of the world, and it also plays a key role in the learning process.

Intelligence has been defined in many different ways- including, but not limited to abstract thought, understanding, self-awareness, communication, reasoning, learning, having emotional knowledge, retaining, planning, and problem solving. Intelligence is most widely studied in humans, but has also been observed in animals and plants. Artificial intelligence is the simulation of intelligence in machines.

Perception is the organization, identification, and interpretation of sensory information in order to fabricate a mental representation

Reason is the capacity human beings have to make sense of things, to establish and verify facts, and to change or justify practices, institutions, and beliefs.  It is closely associated with such characteristically human activities as philosophy, science, language, mathematics, and art, and is normally considered to be a definitive characteristic of human nature. The concept of reason is sometimes referred to as rationality and sometimes as discursive reason, in opposition to intuitive reason.  Reason or “reasoning” is associated with thinking, cognition, and intellect. Reason, like habit or intuition, is one of the ways by which thinking comes from one idea to a related idea. For example, it is the means by which rational beings understand themselves to think about cause and effect, truth and falsehood, and what is good or bad.

Dave,

May I kindly request an edit…

Reference: http://www.chatbots.org/ai_zone/viewreply/10723/

EDIT:

“It’s a very interesting”

SHOULD BE:

“It’s very interesting”

And then delete this post.

Thanks!

8PLA • NET - Jul 22, 2012:

Please go on about analogue computer programming… It’s a very interesting and useful for studying ancient computer programming, fundamentally unchanged for millennia.

Yesterday I came across this article, quite an interesting read:

http://secretinfogarden.blogspot.nl/2008/02/programmable-robot-of-ancient-greece.html

Victor said, “IF-THEN rules and pounding a square into a wheel”

An interesting fact, mentioned by one of my college professors, is that technically, a square is a circle with four sides. One of the most difficult aspects of A.I. may be implementing a natural language processor considering the complexity of human language. 

The approach Victor is taking for natural language processing is of course perfectly valid in one field of artificial intelligence known as NLP for short. Yet, implementing a grammar parser on a computer for Subject, Verb, Object is a challenging feat, but the results are elegant because we may relate to that, as how we use grammar as humans.

8PLA • NET - Jul 22, 2012:

One of the most difficult aspects of A.I. may be implementing a natural language processor based on the complexity of human language.

When you approach the problem of language from the perspective of ‘conceptual knowledge representation’, it’s actually pretty simple to build a communication-layer on top of that, that’s capable of inferring grammatical rules based on statistical word-pairing. My system uses this approach, based on existing research (there is quite a lot of research been done in this area).

- http://en.wikipedia.org/wiki/Conceptual_dependency_theory
- http://www.cs.umass.edu/~ronb/papers/bigrams.pdf
- http://scholarwiki.indiana.edu/wiki/index.php?title=Automatic_computation_of_semantic_proximity_using_taxonomic_knowledge