I'm not sure if this is the best place to post this but it should be ok. I've been thinking, how is the human brain capable of pulling a (seemingly?) random number out of the air? Would it be possible to write a program that could generate random numbers (in a similar way to how the human brain might do it) without using a psuedo random number generator (e.g. rand())? I don't know an awful lot about A.I. but I've been thinking about this problem for a while and thought I would post it here and see what other people's (people with more experience than me in A.I!) thoughts and views were. I'll let this post sit while I go to college, do an exam and a couple ICT lessons and see what I'm left with when I get home! Please post your thoughts and ideas on this topic. Cheers!

Humans are very bad random number generators. They tend to not vary their numbers a lot and do not handle repetition in a random way. In fact there has been programs that 'predict' the answer of a human in the case of binary digit generation with a score significantly higher than the statistical average of 50%.

If you need randomness, don't look into AI.

If you want to make a human-like random number generator, I think the best you can do is train a markov chain with real human input and then use it to generate human-like output.

[edit]

You maybe interested in this page :
http://wetzel.psych.rhodes.edu/random/intro.html

Quote: Original post by InsaneBoarder234
[...]how is the human brain capable of pulling a (seemingly?) random number out of the air?[...]

It isn't. People are _TERRIBLE_ at coming up with random data. Psuedorandom number generators are far better sources of seemingly random numbers than a person calling numbers out. Sure, with pencil and paper you could apply a psuedorandom number algorithm yourself (adding in the tiny bit of randomness humans are good for), but I don't think that is what you're talking about.

Maybe this is why psychics are still in business?

People pick 7 and 3 way too much. Unless you tell them people pick 7 and 3 way too much...

Quote: Original post by Yvanhoe
You maybe interested in this page :
http://wetzel.psych.rhodes.edu/random/intro.html

That was an interesting read.

Well, previously I thought that the human brain was good at picking random numbers but after today I strongly agree that it isn't. While walking to college I decided to try and think of random numbers in the range of 0-50, just reeling off loads of "random" numbers but I noticed patterns (well sort of, either way I was bad at it).

After reading that page though I now think along the lines of "is anything truely random?". That page says that if nothing was random then we would be able to predict everything but I don't agree. Having only just finished reading it I haven't given it an awful lot of thought but at the moment I cannot think of anything that is truely random. I now believe that everything happens for a reason but the reason we cannot predict everything is because there are so many factors that go into everything that people are unable to take them all into account.

Well I guess thats the end of that idea eh :)

Cheers all!

There is some small amount of debate over the existence of randomness in quantum physics. Quantum physics successfully explains things at the atomic level as probabilities. The location of electrons is explained using probability distributions. A radioactive atom has a certain probability of decaying. The question is: are these events random, or are there hidden variables that we have not been able to measure? There are some experiments that show that hidden variables cannot explain what is observed, but there are some who debate the validity of the experiments.

Either way, at the scale we normally experience things at, there are just too many variables to precisely predict things. Lots of things have chaotic behavior, where a small difference in initial conditions can drastically affect what happens. Modern weather prediction is done using powerful computers that run complicated simulations, and the initial conditions are varied slightly to try to find patterns that occur even with the slight variations.

random numbers is stupid but your concept is good look i am not a programmer but i am trying to learn and i think that a random AI would make games different and better more real life ??? something like randomized paths for AI gaurds and random ways an AI can ask you to acomplish a task or quest is this possible

Quote: Original post by Anonymous Poster
random numbers is stupid but your concept is good look i am not a programmer but i am trying to learn and i think that a random AI would make games different and better more real life ??? something like randomized paths for AI gaurds and random ways an AI can ask you to acomplish a task or quest is this possible

Your friend at the end

A computer algorithm with a certain input will *always* have the same output (assuming the hardware is working right...)

So the *only* way to make a random number generator that doesnt always output the same series is to give it different inputs (usually the input is the current time).
I heard some machines (encryptors) have hardware random (for example based on the radio radiation around) so they achive greater randomness. Again this isnt real random generator algorithm, only random input for it.

So how does the brain output random numbers? who says they are? maybe if you could reset the brain back to the previous moment it will output again the same number (so its not generating anything random). Another option would be that the brain has true random numbers using the randomness that (suppposbly) exists in the quantom level, which is possible since the brain is a nanotechnolagy chemical machine. In the first case it works the same way as AI (produces a seemingly unpredictable number based on memories and other inputs), in the second case no AI will be similar since computers dont have true quantom randomness.

Random number generation is a very interesting problem, but I think it has nothing to do with AI, only with finding a way to get random input (so its a problem for physists not programmers).

The algorithmic side of the problem is to make a very big random imitating sequence with very small random input. But again this not a problem for AI programmers, but rather to a mathematician.


Iftah.