[lrh9]

So essentially semantic understanding is understand the meaning of data?

I think I need to look at more of your posts to learn about how you propose to imbue machines with semantic understanding.

[TrueAndroids]

So essentially semantic understanding is understand the meaning of data?

Right and maybe more specifically

 "... the derivation and matching of the semantics of computational content to that of naturally expressed user intentions

in order to retrieve, manage, manipulate or even create content,

where "content" maybe anything including video, audio, text, processes, services, hardware, networks, etc."

FROM:  http://www.ieee-icsc.org/

[trotter]

Often questions here are devoted to the pursuit of the development of a.i. However, what would we do if one day after all of the book reading, Internet searches, forum discussions, and programming, we finally had something we considered to be a release candidate for a.i., how would we test it and optionally why would we use that test?

Intelligence tests such as the IQ test come to mind.

The Turing Test. I wouldn't send it to the Turing Test with the goal of deceiving the judges. I think that if it is truly intelligent, it should be able to convince them it is intelligent in discussion.

You can ask your question differently : What make you think that the people you talk in your IM are intelligent ? Did you have to give IQ tests to your contacts ? How about people IRL ?

Turing proposition is basically : "if it appears to be intelligent, then it's intelligent."
It seems stupid, but that's logical. You are the one who judge if your contact is intelligent or not.
It's the same as "If this flower appears to be pretty, then it's pretty."


Like Art I'm thinking of a test protocol to detect chatterbot. Here we go :

1)Can the test subject remember my name ?
It seems simple to implement but it's actually a lot of work :
"Bot: What is your name ?
User: sorry brb.
Bot: Hello, nice to meet you sorry brb.

"Bot: What is your name ?
User: it's Bob you curious little fox.
Bot: Hello, nice to meet you Bob you curious little fox."

2)Verify if the contact can link answers with context. You can do that simply with saying "and you ?" or "why ?".

3)Check if long sentences can be understood.
"user: He say you are pretty, I say you're stupid.
Bot: You think I'm pretty ? Thank you !"

4)Like the name, try to see if secondary infos (place of residence, health, favorite food) are recorded.

5)Check if logical links can be made. "Rex is a dog. A dog have four legs. How many legs have Rex ?"

[lrh9]

An intelligent agent can appear unintelligent, and an unintelligent agent can appear intelligent.

An intelligence metric based on the appearance of intelligence is unreliable.

I'd modify the Turing test to this: 'If an agent can pass a test then it is intelligent for that test.'

The people I converse with and interact with pass conversation and interaction tests, therefore they are intelligent in regards to conversation and interaction. The same people might not be able to pass tax handling tests, therefore they are not intelligent in regards to taxes. For help with my taxes I would go to an agent that had passed tax handling tests.

[Art]

Conversely, the majority of people in the world are perhaps only intelligent in a very small area of expertise which would make the world, as a whole, fairly unintelligent.

Hmmm...now we know.

If a passing grade in class was still a D (based on A=highest and E/F = Failing), and your medical doctor received a D+, would you want him / her treating you?

It comes down to expertise in a given area. Interesting thoughts Larry.

[lrh9]

I would prefer a doctor with a higher grade treated me instead of a doctor with a lower one.

Sometimes there is no choice though. The world - as a whole - has to make do with what we have because that is all we have.

However, often grading systems reflect more than mastery of test content, but also other metatests such as effort, presence, and consistent test completion.

I'd be less worried about a doctor who had received a D because he or she skipped class but was otherwise a medical genius. I'd be more worried about a doctor who worked hard but had difficulty mastering test content.

Of course, this is all hypothetical. Even if a person passed a medical course he or she might not be a doctor. I think there are certain thresholds governments and medical organizations impose on medical employees because the margin of error when it comes to human lives should be as small as reasonably possible.

That is a principle we might want to remember when considering whether or not to assign artificially intelligent agents to tests.

[Art]

Some good points, Larry.

[Data]

What is intelligence?

It cant be the ability to recall something learned earlier, a computer is good at that but not intelligent, so simply being able to give the right answer to a question, is not going to prove anything.

For me intelligence is the ability to solve problems and a big part of that is creativity, Einstein for example would “dream up” theories and put them to the test using maths, this part of the mind is our unconscious, the place where dreams come from, can we ever really expect a computer to be able to do the same, we don’t even fully understand our own minds, let alone build one.

You know the more I think about it the less I am convinced we will be able to do it.  

We = Human race

EDIT:
PS ive been following this topic, interesting.

[lrh9]

so simply being able to give the right answer to a question, is not going to prove anything.

For me intelligence is the ability to solve problems

Isn't a problem just a real world question, and isn't a solution just a real world answer?

[Data]

Isn't a problem just a real world question, and isn't a solution just a real world answer?

Yes Irh9 when you put it that way I have to agree with you.

I will try to elaborate on my thoughts and maybe help a little to this topic.

All through Mans history there have been times that he has made big leaps forward, one good example that fits what I am trying to say is Sir Isaac Newton and gravity, he not only started to solve the problem of gravity, he also realised there was a problem. It’s that spark that we humans have, the closest I can relate it to is creativity.

If we simply make a test for a computer to pass, by asking it questions and it giving the correct answers then all the programmer has to do is make the computer give the correct answer to the question, but is that intelligence?

People say mathematicians are intelligent and well they are, aren’t they? But a calculator could do the same kind of thing maybe even quicker but again is not intelligent.

This is all good to ponder over, now let’s take this as an example. 

Computer: sorry insufficient data to answer your question.

Human: dreams up theories in its mind to find the missing data.

Call it what you will, some may say it’s the soul; some call it psyche, I think of it as creativity.

Can a computer ever think for itself? Let’s hope so.

Can a computer pass a test and give the right answer? If all it has to do is answer questions then yes why not.

Google already answers most of my questions. 

Can a computer come up with the question? Now there is a thought. 

[lrh9]

"Necessity is the mother of invention."

Isn't the motivation for all human questioning and creativity desire or necessity?

I don't want artificial intelligence to desire, but all computers have a characteristic similar to having a need: they must operate according to their programming.

I think a correctly designed and implemented artificial intelligence could be metaprogrammed to strive to understand and create.

If there will be any barrier to artificially intelligent creation it will be the nature of understanding and creativity: whether or not they are algorithmic processes and, if they are algorithmic, whether or not they are too complex.

[Art]

The calculator or computer is just a tool much like a saw, hammer and square.

It is the accompanying knowledge / intelligence of the individual, be it mathematician or
carpenter, that defines the applied intelligence.

Having a magic wand does not make one a sorcerer. As with most things, it takes wisdom,
intelligence, know how....

If we construct a chatbot that can act on it's own providing knowledge, correct answers and generally
helpful insights to problems, then perhaps we could say that it is demonstrating a degree of intelligence.

Without the programming knowledge or database...it's just hardware and software.

[Data]

If there will be any barrier to artificially intelligent creation it will be the nature of understanding and creativity: ...

Im glad you agree with me Irh9. It wasnt just me thinking that.

Art, I guess we could call it intelligence, honestly I don’t exactly know what Intelligence really is, sounds odd but when I really think about it there seems to be a number of things going on and we bundle them all together and call it intelligence.

If we can accept Ai as being different to human intelligence, I do have a little trouble with that, but never the less it has to be a form of intelligence. I think 

[trotter]

I like what lrh9 said, "Necessity is the mother of invention."

We humans are driven by our needs. If we only had logical abilities but no urge to use it, we'll be like robots, right ?
If someone had his needs solved magically right after he wants something, does he need to talk ? I don't think so...

That's why I think the answer may be to make a program with some basic needs in a virtual world, like a virtual pet where we can interact with.

(Maybe I'm a bit off-topic here, sorry : )
Then give the program the abilities to detect stimulus and link it with response. For instance, every time we give him food, we can write "here is your food", "time to eat", "you must be hungry, have some food".
The bot can then link the stimulus (the word "food") with action of eating (Pavlov anyone ? Since I don't have a clue in cognitive science, I'm stuck with Behaviorism approch ). Once we have a word (or action) linked to a need, if the bot have this need, we use the stimulus to express it (the bot say the word connected with eating="food").
This is an example with food but we can use anything in Maslow's hierarchy of needs. We can also use tons of stimulus (leading to pleasure/pain).