Kylotan says exactly what I wanted to say.. What if human intelligence could actually be understood, would humans still be deemed intelligent or would that discredit them.
Perhaps 'measuring' intelligence is not such a one-dimensional objective. How intelligent a system is should be replied to with a program. As in, how intelligent is system A. Reply: "In case a happens, it replies with ..., ..."
Oh and this statement has helped me (though regarding a different aspect of the problem): Asking if a machine can think is just like asking whether a submarine can swim.
What is really AI?
Quote:Original post by arithmaOh and this statement has helped me (though regarding a different aspect of the problem): Asking if a machine can think is just like asking whether a submarine can swim.
Oh, I like that. Very good way of looking at it.
As for saying that "you can argue that is 'not intelligence', because you can see all the inner workings and predict it with a high degree of accuracy". People used to think that Volcanos and Nature and the Sun were intelligent because the behaviour they displayed appeared to be intelligent.
Now obviously, what we know now proves they are not intelligent.
If it were proved that human beings are deterministic machines, 100% predictable, then I'd have to say we're not intelligent either, we'd be, well, nothing more than deterministic machines ;)
Quote:Original post by Hinkar
If it were proved that human beings are deterministic machines, 100% predictable, then I'd have to say we're not intelligent either, we'd be, well, nothing more than deterministic machines ;)
So by your definition, something needs to be stochastic (display random behaviour) for it to be intelligent. Most credible researchers in psychology, neurology, philosphy, etc., would disagree strongly with you.
It appears that my first comment was overlooked as merely a trite response... in fact, it was a very specific point... most behaviours displayed by humans are merely complex behaviours that can be explained in terms of more fundamental behaviours and basic physiological processes, yet most people think we are displaying "intelligent behaviour".
For example:
- the inate ability of humans to quickly identify objects as either human faces or not is nothing more than hard-wired pattern recognition system specialised on characteristics of human faces.
- our ability to learn a language is based on the rote learning of associations between symbols and objects... as children, when our brains are still young, we can encode those associations in specific neural pathways near the junction of the temporal and occipital lobes of the cortex, close to the regions where speech is controlled... as adults we can't do that and we rely on our hippocampus and other temporal lobe structures to encode this information... and thus find it harder to learn to speak foreign languages. Either way, it's just data storage and retrieval. Nothing intelligent there, just complex cognitive behaviour.
The best possibility for intelligent behaviour in humans lies in our higher order planning and prediction systems that are mostly found in our frontal lobes. Again though, a lot of good research has shed light on the fact that much of what goes on in the frontal lobe is just rapid solutions to simple planning and prediction problems, where the results are used to modulate lower order systems (such as speech and motor (muscle) control). So where then is the "intelligence" in a human?
From my perspective, there isn't any, unless you accept that "intelligence" is a (poorly chosen) term that can be used to describe a spectrum of complexity of behaviours. Some systems are therefore more intelligent than others because they display more complex behaviours (or in other words, behaviours originating from more complex heirarchies of lower order behaviours).
Can we do this in a computer? Absolutely. Is it human intelligence (or its equivalent)? Only if it experentially and behaviourally equivalent.
Anyway, that's my two cents (and more)... getting off my soap-box now... ;)
Cheers,
Timkin
There's a reason why 'artificial' plays a part in the term. It's just an emulation of intelligence, not the creation of it. We think humans and animals have a certain degree of intelligence (I know plenty of the former who do not exhibit any at all), therefore we try to design computer systems that exhibit similar and slightly unpredictable behaviours. Intelligence, in real terms, is an organism's ability to survive, I would say.
Intelligence is what it is because we said it was that. Intelligence is just a word.
Intelligence is what it is because we said it was that. Intelligence is just a word.
Quote:Original post by NickHighIQ
therefore we try to design computer systems that exhibit similar and slightly unpredictable behaviours.
Again I reiterate that people working in AI research (and I am one of these people) are not, in general, trying to create systems that are unpredictable. On the contrary, you cannot release artificial systems into the real world IF they are unpredictable. Predictability (or a lack of it) has NOTHING to do with intelligence. Intelligent systems are NOT unpredictable.
Intelligence is what it is because we said it was that. Intelligence is just a word.
'Intelligence' may just be a word, but every used word in a vocabularly has meaning. The problem with this word is that people cannot agree on its meaning. It's really the thing that 'intelligence' refers to that is the subject of the debate, rather than the word used to label that thing.
Personally?
I would love to see the birth of AI, and I say birth for these reasons:
1. An entity which does not understand that it has place and being is, according to some, not intelligent enough to be AI. I agree with this, as every person understands that in some way, they are a being.
2. The AI does not only have a system which finds out different solution systems to problems, but their is a system which allows it change its learning system. For example, when you or I learn how to read, we can begin inputting information through textual reference, instead of audible/symbolic input. An AI which cannot adapt a new, more efficient system is unable to gain conceptual intelligence, and remains at a fairly stationary IQ, as programmed by the creator.
Programming this kind of thing is most definitely difficult, as it is very ambiguous. But, I believe it is the way to create an AI equal to ourselves, or perhaps more intellectually capable than us.
I would love to see the birth of AI, and I say birth for these reasons:
1. An entity which does not understand that it has place and being is, according to some, not intelligent enough to be AI. I agree with this, as every person understands that in some way, they are a being.
2. The AI does not only have a system which finds out different solution systems to problems, but their is a system which allows it change its learning system. For example, when you or I learn how to read, we can begin inputting information through textual reference, instead of audible/symbolic input. An AI which cannot adapt a new, more efficient system is unable to gain conceptual intelligence, and remains at a fairly stationary IQ, as programmed by the creator.
Programming this kind of thing is most definitely difficult, as it is very ambiguous. But, I believe it is the way to create an AI equal to ourselves, or perhaps more intellectually capable than us.
Quote:Original post by dbzprogrammer
1. An entity which does not understand that it has place and being is, according to some, not intelligent enough to be AI. I agree with this, as every person understands that in some way, they are a being.
The problem is that no amount of evidence will convince a skeptic judge that a machine understands anything. See the comment about swimming submarines.
Quote:2. The AI does not only have a system which finds out different solution systems to problems, but their is a system which allows it change its learning system. For example, when you or I learn how to read, we can begin inputting information through textual reference, instead of audible/symbolic input. An AI which cannot adapt a new, more efficient system is unable to gain conceptual intelligence, and remains at a fairly stationary IQ, as programmed by the creator.
I don't think this property is central to the idea of intelligence. If an entity starts out with all the ways of learning that you can think of already active, does that count against it?
Quote:Programming this kind of thing is most definitely difficult, as it is very ambiguous. But, I believe it is the way to create an AI equal to ourselves, or perhaps more intellectually capable than us.
You didn't describe any "way to create" anything. You just gave two reasons for saying "birth" (although I didn't follow the argument), which may pretend to be criteria for something to be intelligent, one completely untestable and one ambiguous and IMO mostly irrelevant.
My personal take is that the aim of AI is creating artificial entities that can solve problems. I don't think there will be a "birth" of any kind. We'll just get better and better at making those entities, and they will be able to solve more and more problems. "Learning" is probably a good way to get to solve more problems, but I wouldn't make it a requirement for intelligence.
Quote:Original post by TimkinQuote:Original post by Hinkar
If it were proved that human beings are deterministic machines, 100% predictable, then I'd have to say we're not intelligent either, we'd be, well, nothing more than deterministic machines ;)
So by your definition, something needs to be stochastic (display random behaviour) for it to be intelligent. Most credible researchers in psychology, neurology, philosphy, etc., would disagree strongly with you.
It appears that my first comment was overlooked as merely a trite response... in fact, it was a very specific point... most behaviours displayed by humans are merely complex behaviours that can be explained in terms of more fundamental behaviours and basic physiological processes, yet most people think we are displaying "intelligent behaviour".
For example:
- the inate ability of humans to quickly identify objects as either human faces or not is nothing more than hard-wired pattern recognition system specialised on characteristics of human faces.
- our ability to learn a language is based on the rote learning of associations between symbols and objects... as children, when our brains are still young, we can encode those associations in specific neural pathways near the junction of the temporal and occipital lobes of the cortex, close to the regions where speech is controlled... as adults we can't do that and we rely on our hippocampus and other temporal lobe structures to encode this information... and thus find it harder to learn to speak foreign languages. Either way, it's just data storage and retrieval. Nothing intelligent there, just complex cognitive behaviour.
The best possibility for intelligent behaviour in humans lies in our higher order planning and prediction systems that are mostly found in our frontal lobes. Again though, a lot of good research has shed light on the fact that much of what goes on in the frontal lobe is just rapid solutions to simple planning and prediction problems, where the results are used to modulate lower order systems (such as speech and motor (muscle) control). So where then is the "intelligence" in a human?
From my perspective, there isn't any, unless you accept that "intelligence" is a (poorly chosen) term that can be used to describe a spectrum of complexity of behaviours. Some systems are therefore more intelligent than others because they display more complex behaviours (or in other words, behaviours originating from more complex heirarchies of lower order behaviours).
Can we do this in a computer? Absolutely. Is it human intelligence (or its equivalent)? Only if it experentially and behaviourally equivalent.
Anyway, that's my two cents (and more)... getting off my soap-box now... ;)
Cheers,
Timkin
That's pretty much it.
Author
Quote:Original post by Rixter
AI is search.
what about onthology..relational informations memorized with observational atoms
This topic is closed to new replies.
Advertisement
Popular Topics
Advertisement
