Artificial intelligence, the decision way

The worldwide attempt to construct artificial intelligence has generally been acknowledged as a failure. I would like to discuss constructing artificial intelligence based mainly on the notion of decision. A decision I define as a point where something goes one way in stead of another. The only place this possibly occurs in a computer is in the random function (or by user input, or error). So the question becomes, what programmatic constructs can be made of the random function? The goal is of course, to make constructs that are equivalent to emotions.I'm supposing the randomfunction only returns 0 or 1.Systematically:
simple - call random
selfsustaining - do while call random<>0
spawning - do while call random<>0
call selfsustaininghrmmmm what more? I don't know...In any case try to get a very big complex thing running, by decision alone.Next is to let this complex thing manipulate something. Like for instance language, or pixelpainting.The problem here is cognition. Let's suppose that the future is real, consisting of chances, and that the decisions are made in connection to these chances. That is already cognition of a sort, the connection between the chances and the decision.Then what?I don't know, but this making of a complex of structures that act merely by volition sure does seem to be the way to go, to construct artificial intelligence.Let's all think about decision, so we can come to know that creation is true.regards,
Mohammad Nor SyamsuThis message has been edited by Syamsu, 04-13-2005 01:57 AM

Syamsu, you almost could have an entire topic with yourself. You answer many of your own questions! Most of us can't quite flow with the words that you are choosing to use. Here is an idea.Pretend that you need to discuss your idea of artificial intelligence with a room of fifth grade students.Think how you will explain your theory to them. Use easier words and try and connect your ideas so that these eager young students will understand you. (As an example, explain the "notion of decision" so as a fifth grade student would understand you. Use a parable or a story to make the point clear.Regards,
PBBring the topic back and show us how you have made it easier to read.

Well this topic is of course especially for those who have a handle on programming, altough programming is easy basic logic which anyone can do in a minute. I should also point out that noone on this forum is actually defining decision, or like term, in a workable way. I would be happy to use a term defined in the glossary of the site, but what I believe to be the main thing in the concept of intelligence, decision, is not in the glossary. So you may consider that you not kill of discussion on the subject of decision altogether, if you not allow this, seems to me, rather simple and straightforward definition of decision. And by the way, the randomfunction is the most fascinating thing in the computer, which any selfrespecting nerd could not possibly neglect to have an opinion on. I mean this topic should have general "supportive" interest, also of evolutionists, eventhough I predict it leads towards creationism.----The worldwide attempt to construct artificial intelligence has generally been acknowledged as a failure. I would like to discuss constructing artificial intelligence based mainly on the notion of decision. A decision I define as a point where something goes one way in stead of another.For instance the biologist Gould theorized that if time were wound back, that evolution may turn out differently. As for example a comet may not have struck the earth, and the dinosaurs would still roam the earth and evolve into other species. The point at which it was determined that the comet would strike the earth, in stead of not striking it, I call decision.The only place decision possibly occurs in a computer is in the random function (or by user input, or error). When the random function is called the computer generates a number between zero and one. Per time the probability that the computer generates a number that is closer to one is equal to the probability that it is closer to zero. The rest of the computer behaves according to rules of calculation, and the outcome is predetermined according to the rules. So having identified the random function as the source of decision in a computer, the question becomes, what programmatic constructs can be made of the random function? The goal is of course, to make constructs that are equivalent to emotions, rather then concepts of intelligence based around calculation.For sake of simplicity I'm supposing the randomfunction only returns 0 or 1.A systematic overview of the possible ways in which the random function can be used:- simple
call random
(the random function is called, and that's all)- selfsustaining
do while call random<>0
(the random function is called and depending on the result, it is called again and again, until the result is zero)- spawning
do while call random<>0
call selfsustaining
(the randomfunction is called and depending on the result another instance of the "selfsustaining" procedure is called, so that many instances of the selfsustaining procedure may run at once)- consistency
do while call random=call random
(the random function is called as long as the outcome is equal to the previous call of the random function)- inconsistency
do while call random<>call random
(the random function is called as long as the outcome is not equal to the previous call of the random function)hrmmmm what more? I don't know... Please suggest any other fundamentally different calls to the random function.In any case the goal is to try to get a very big complex thing running, by decision alone.Next is to let this complex thing manipulate something. Like for instance language, or pixelpainting.The problem here is cognition. Let's suppose that the future is real, consisting of chances, and that the decisions are made in connection to these chances. That is already cognition of a sort, the connection between the chances and the decision. At the point of decision the possible outcomes 0 and 1 are "known", there is cognizance there of the possible states. So what is manipulated first are the chances of a pixel being painted, and afterwards the pixel is actually painted.Then what?I don't know, but this making of a complex of structures that act merely by volition sure does seem to be the way to go, to construct artificial intelligence.Let's all think about decision, so we can come to know that creation is true.regards,
Mohammad Nor Syamsu

It seems to take a long time...... to get this post accepted/rejected.regards,
Mohammad Nor Syamsu

Same old attempt to rename chance as decision.Forget it.

---

This is in response to Syamsu's request to have this topic reconsidered. But Syamsu, this is a discussion forum. If nobody agrees with the foundation of your argument, then you have nobody to discuss your new ideas with. Nobody wants to talk about the application of an idea that they find fundamentally flawed. This has nothing to do with being right or wrong. It has everything to do with how a discussion works. I understand. Let me repeat once more. Nobody else is willing to use your working definition. Nobody else is willing to go into practical applications. Thus, this topic will generate only two things:1. Arguments about what decision is.
2. SilenceNeither of these is any grounds for a new topic. Maybe. But there's no basis for discussion with anybody here. If you want to pursue it, then work out the issues brought up in the "decisions" thread. If you can address those issues and find other people who are willing to share your working definition of "decision," then we can revisit this PNT. If you're unwilling to do that, then there's no reason to start another thread which will generate exactly the same responses.Thanks.

---

'Nuff said.Thanks MT... in that "'XYZ-PDQ' / 'thanks for letting me know my zipper was down'" kind of way. Closing this down.