The Turing test

How does science work? And what's all this about quantum mechanics?

Moderators: AMod, iMod

Post Reply
User avatar
Aetixintro
Posts: 319
Joined: Tue Jan 08, 2008 7:44 pm
Contact:

The Turing test

Post by Aetixintro » Tue Jun 01, 2010 4:22 pm

I'm curious as to what the Turing test is good for.

On the one hand you have this computer set-up that's programmed semantically by a number of programmers. On the other you have this semantic (naturally) one person. Between these two, the computer and the person, another person is asked if this person can separate the machine from the human, the person.

So if one considers this from the two year old to the hundred year old, being the span of the age of this person who is to determine the person from the computer, the older person naturally has a greater range of questions to ask than the younger person, usually, because age means greater reflection.
Naturally, the older person does this better than the younger (you may deduct the very oldest).

If you in addition to this give a time limit for making the question, like 10 minutes, I think it's reasonable to say that this gets severely much more difficult. So if one is combining the very young age with a relative short time-limit to ask the questions, I think it's fair to suggest that a considerable number of people will fail the test, being unable to separate the person from the semantically programmed computer if it's comprehensive enough.

So what is the Turing test supposed to prove? What is its significance? The person and the programmers are semantically indistinguishable. That is, the computer system may even prove more advanced than the person if one puts enough programming into it. Sure, one can point to the possibility for actually being able to present a computer system that can answer like a person can or even better, but this is already done and is surely matching the younger people in combination with the time-limit.

My conclusion is that the Turing test is of no significance (anymore) since the programming possibilities are so vast!

What do you think? :)

Impenitent
Posts: 2293
Joined: Wed Feb 10, 2010 2:04 pm

Re: The Turing test

Post by Impenitent » Wed Jun 02, 2010 12:32 pm

think about the test again...

it doesn't prove that the machine thinks or has intelligence

it is all a question of what you believe and how you react to that belief...

If you don't think that box of switches has sentience, why are you conversing with it?

-Imp

mickthinks
Posts: 782
Joined: Thu Oct 18, 2007 1:10 am
Location: Augsburg

Re: The Turing test

Post by mickthinks » Wed Jun 02, 2010 2:43 pm

You raise some good points, Aetixintro. As I understand it (which is at a fairly basic level) Turing's test was never intended to be a practical test and probably can never be so. It's best thought of as a thought experiment, addressed to those who question whether an information-processing machine could be really conscious. Imagine a skeptic saying something like "we can see the transistors, we understand the mindlessness of the electronics and the logic, so we know this isn't real conscious thought". Turing responded with this famous thought experiment, which suggests in effect, if you cannot tell whether you are talking with a human being or a machine then you have no reason to differentiate between the thinking that each is doing. It's not a test so much as an argument that artificial consciousness is possible.

Richard Baron
Posts: 204
Joined: Wed Oct 17, 2007 6:55 am
Contact:

Re: The Turing test

Post by Richard Baron » Wed Jun 02, 2010 5:37 pm

It is worth having a look at Turing's 1950 paper, which is available here:

http://www.abelard.org/turpap/turpap.php

His emphasis is on whether machines could think, rather than on whether they could be conscious. it is, however, a very interesting question whether thinking implies consciousness, or consciousness implies thinking, or both, or neither.

Did Turing think of his proposal as a purely hypothetical thought experiment, or as something that might come to pass? I incline to the latter. In the second paragraph of section 6, he says: "I believe that at the end of the century the use of words and general educated opinion will have altered so much that one will be able to speak of machines thinking without expecting to be contradicted". This follows remarks about the likely increase in computer power.

He was a bit optimistic on timing. I think there is an annual competition to see whether a computer can pass the test, and none has yet succeeded.

He also did not give much consideration to computers that might be better models for the brain than digital computers as they were then conceived. This does not undermine his argument. There is no obvious need to imitate the brain's way of doing things in order to imitate its outputs, and the concept of a Turing machine covers all of the well-behaved self-contained digital computers we can so far think of. But the point may colour our reading of his work. It was not long afterwards that John von Neumann explored such issues in The Computer and the Brain. And there has been plenty of more recent work on neural nets and on the relevance of analogue computing.

Should we read Turing's proposal as a purely hypothetical thought experiment, which helps to clarify our notion of thinking? It certainly does that. But my feeling is that it won't be long before Hal is wondering how to tell whether human beings can think.

jetsetjason
Posts: 107
Joined: Wed Oct 17, 2007 1:39 pm

Re: The Turing test

Post by jetsetjason » Wed Jun 02, 2010 9:07 pm

Oddly enough, I have just been reading a book about Turing ( cor, you know when you belong eh... )

http://www.amazon.com/Man-Who-Knew-Too- ... 0393052362

I think, from the context, of things, he was just speculating, as the media, at the time, liked the idea of thinking machines and being all sci-fi like.

What is interesting, is that, AI research has pretty much failed BUT as the creators of 'the sims' and other games have found, it is far easier to 'fake intelligence' than create it.

Computers are tools, nothing else.

On a point of interest, Alan knew Wittgenstein and was a member of the moral science club.

I found the information on Alonzo Church most interesting, he seemed rather eccentric, as was Turing.

COOL

User avatar
Psychonaut
Posts: 465
Joined: Wed Oct 17, 2007 10:40 pm
Location: Merseyside, UK

Re: The Turing test

Post by Psychonaut » Thu Jun 03, 2010 3:44 am

I wonder whether they do a control, and how often a human gets labeled as a program.

User avatar
Aetixintro
Posts: 319
Joined: Tue Jan 08, 2008 7:44 pm
Contact:

Re: The Turing test

Post by Aetixintro » Thu Jun 03, 2010 11:41 pm

Acknowledged, everybody! Thanks for your replies!

If you sense/get the notion that I'm missing something by not replying to you in proper manner by this, please tell me!

Cheers! :)

Post Reply

Who is online

Users browsing this forum: No registered users and 4 guests