4 replies to this topic

[greenpeace]

Today a friends asked me to say a number over the phone. I asked him why but he insisted me just say a number, that's all. I told him one 5 digit number and he disconnected the phone. Till now, I am not sure why he asked for this number and what was the number I have given to him... but this raised another question in my mind.

Is there any relation between a random number generated by a human brain and a computer? Which one is actually random, one I told him without any thought or a number generated by a computer program?

When I was in college, I learned about programming functions like x=rand() where a compute generates a random number using a built-in random function and assigns this random value to the variable "x". But internally this random functions must work of certain variables so that a computer can create a random number, I think. It can be anything - time, date or any available parameter - twisted and coded in such a way that it becomes a good random. But it always requires an absolute value, at a particular moment, to generate a random. In short, eventhough we may find a compute generated random is absolutely random, it is not so internally.

But what about we human beings? When someone asks you to say a number, we probably link this to our date of birth, today's date or present time and create a random out of it. We may also modify a number displayed nearby, like a car number plate, make a quick alteration and tell that number as random. But what it we do not depend on any of these tricks and produce an absolutely random number out of our mind? Can't we do that?

Which one these numbers, machine generated or human generated, is more random? Is this one of the areas in artificial intelligence where human brain outperforms super computers?

(I will call my friend later and tell you why did he asked for that random number, the "Newton's Apple" )

[shadowx]

Neither is more random than the other. At the splitsecond the command is processed to make a random nmumber (Be that a signal in your brain or the RAND command in a computer program) a super computer could calculate, 100% the number that is produced if it new the variables. for example if your brain has a tendency toward a number, lets say 27, and a sequence, lets say n, n+2, n-7, n+5 (so lets say N is the first number in a sequence and your brain chooses the number 27 then the sequence is: 27, 29, 20, 32 (27292032)) then if the super computer knows these tendencies it will KNOW which number you are going to choose, not just guess, Same goes for a computer generated code. If the super PC knows the timestamp being used and the algorithm then its not random so long as the super computer knows the exact time the command is issued.

A better way to generate random numbers is to use the background radiation that is all around us. By counting the background radiation (BR) you could then feed it into a hashing algorithm (perhaps as simple as an MD5 hash) and get a more random number. If you then combined that with a random number found by using a hash of a timestamp and then a random number found by generating a random number from..... the earths magnetic field and then joined them together in one giant number and then used a hashing algorithm to bring that huge number down to say 10 digits, thats a fairly random number.

But then you could argue that at the exact moment of the big bang the exact properties (size, weight shape, magnetic field and radioactivity) was decided so that the "random" number was actually determined billions of years ago, and if you could take a super super computer back to the big bang you could actually predict with 100% certainty what the number would be billions of years from now.

But that goes into deep things like is fate real etc? (because lets face it, if there were no life forms on the unverse then every particle and every bit of energy was created and sent out at a certain direction from the big bang, so their paths arent random so in theory if you had a computer clever enough to measure every atom and subatomic particle ejected from the big bang you could actually predict, with 100% certainty where the planets would form, their exact size at any point in time down to the atomic weight etc... because each particle has a pre-determined path.

But with life forms you then introduce variables like a colony of ants moving a small pebble out of their path, where if there were no ants that pebble would be moved according to the interactions with other particles that had a pre-determined "fate" thanks to the big bang.

Still with me?

[greenpeace]

shadowx, on Nov 19 2009, 01:38 PM, said:

Neither is more random than the other. At the splitsecond the command is processed to make a random nmumber (Be that a signal in your brain or the RAND command in a computer program) a super computer could calculate, 100% the number that is produced if it new the variables. for example if your brain has a tendency toward a number, lets say 27, and a sequence, lets say n, n+2, n-7, n+5 (so lets say N is the first number in a sequence and your brain chooses the number 27 then the sequence is: 27, 29, 20, 32 (27292032)) then if the super computer knows these tendencies it will KNOW which number you are going to choose, not just guess, Same goes for a computer generated code. If the super PC knows the timestamp being used and the algorithm then its not random so long as the super computer knows the exact time the command is issued.

A better way to generate random numbers is to use the background radiation that is all around us. By counting the background radiation (BR) you could then feed it into a hashing algorithm (perhaps as simple as an MD5 hash) and get a more random number. If you then combined that with a random number found by using a hash of a timestamp and then a random number found by generating a random number from..... the earths magnetic field and then joined them together in one giant number and then used a hashing algorithm to bring that huge number down to say 10 digits, thats a fairly random number.

But then you could argue that at the exact moment of the big bang the exact properties (size, weight shape, magnetic field and radioactivity) was decided so that the "random" number was actually determined billions of years ago, and if you could take a super super computer back to the big bang you could actually predict with 100% certainty what the number would be billions of years from now.

But that goes into deep things like is fate real etc? (because lets face it, if there were no life forms on the unverse then every particle and every bit of energy was created and sent out at a certain direction from the big bang, so their paths arent random so in theory if you had a computer clever enough to measure every atom and subatomic particle ejected from the big bang you could actually predict, with 100% certainty where the planets would form, their exact size at any point in time down to the atomic weight etc... because each particle has a pre-determined path.

But with life forms you then introduce variables like a colony of ants moving a small pebble out of their path, where if there were no ants that pebble would be moved according to the interactions with other particles that had a pre-determined "fate" thanks to the big bang.

Still with me?

So my doubt "has something in it", shadowx....

CERN is among the two organizations I follow on my Twitter. It is not the time to decide whether everything said in the field of fundamental physics was correct or everything was wrong. Let particle physics work at the LHC and find answers... (Although we are free to make our own philosophies )

From your answer, I assume that a computer starts creating a random number from a variable and this variable can be as complex as the number of protons present in the United States + Mars + an ice cream ! We can increase the degree of randomness by including complex variables, but true at the moment of determination, in to our RAND function like CMBR. Looks good...

And my mind may have an inclination towards a particular number from which it generates a highly random number, thanks to the high-speed neuron circuits of my brain. But I must confess, I was unable to successfully decode how my brain came to this final random. Reverse engineering skills of my brain may not be that effective. But still, this raises another question...
Is my brain much, much sharper than I think of it at the moment? A new field of study for me indeed...

So the conclusion is, according to our present physics, mathematics and neuroscience, an absolutely random number does not exist.. Isn't it?

[shadowx]

it depends how you define the word "random".

For the purposes of say, a board game, a set of dice is random. Its virtually impossible to guess the throw before they stop, however, i expect if you recorded 1 million dice throws using 5 different people (200,000 throws each) i think you would find it isnt really random, but i would imagine it would only be 5% or less biased.

It's like a coin toss, it isnt actually random, it depends on the original orientation of the coin (tails/heads up/down) and the height of the toss. If you tossed a coin 1 million times you'd reasonably expect about 50/50 but that's not what would happen in reality. However it's a lot more random than a computer or your brain could create. Your brain is biased by memory so trying to make a random sequence of 1 and 2 (eg 1,2,1,2,2,1,2,1,1,1,2,1,2) isnt possible because you are *trying* to make it random, so like i did there i had to try to make it *look* random by taking away any obvious pattern and in doing so i make concious decision to put a group of 1's in or a group of 2's so it isnt random anymore, i planned it that way to make it look random.

And a computer is fundamentally limited in that it cannot imagine. A computer can only *calculate* and to do any calculation you need two or more values, in the case of a random number generator these values can be anything from a variable like time and then a multiplier (eg timestamp*257434265235751/3652) or two variables like timestamp and number of keys pressed since RANDOM seconds ago (but of course, RANDOM seconds needs to be based on something too....)

I suppose the big bang was somewhat random as nothing existed to influence it except the "explosion" itself, but then those who believe in a godly creator argue that he/she made the big bang or universe so even that couldntve been random...

I think the closest we can get to truly random would involve some sort of calculation based on a "seed" obtained from single celled organisms. They have no brain so cant be biased, the only influences on them would be environment but provided that was perfectly controlled and the same each time a "seed" was obtained then it could be called almost random. (seed = a number or calculation upon which another calculation is performed to obtain a random number)

The reason is that the particles made at the time of the big bang cant directly influence their behaviour, let's say we had a dish with 10,000 single celled organisms floating freely in water. It is sealed so no air can affect them and no-one touches it. If we divide it in half and say that when we want to get a seed we count the number of organisms on the left side of the dish (they are able to move on their own perhaps...) and use that as our seed then we can assume that the number of organisms is fairly random as they have no brain or other bias for being on the left side. They arent aware of their surroundings and arent affected by their surroundings.

we can then use that random seed to multiply by the current timestamp or current background radiation or whatever and obtain an almost random number. But it still isnt truly random for reasons i dont yet understand!

[greenpeace]

Yes, it depends on how we define the word "Random". I am thinking about the abstract meaning of randomness. For everyday practical applications, there are many random generators.

A coin toss is a good random in conventional standards but when we analyze it in depth, we have to travel back in time, to the point of the origin of the universe from the Big Bang, as it is understood at the moment.

It would be hard for me to believe that "something comes out of nothing"...

Big Bang is a good candidate for the concept of "Absolute Random" due to the non-existence of other factors to decide its fate. Agreed, even though I am neither a fan of the Big Bang theory nor do I expect a "Creator God" beyond the universe. Everyone doubting the Big Bang theory are not expecting a "Creator God", they may be thinking about other possible options. In fact, we are not much familiar with the concept of "Creator God" in India as the religion here does not endorse such a concept with the exception of inferior literature on personal God etc. As any attempt of "study" presupposes ignorance and I am aware that CERN is working to resolve the exciting Physics of cosmic origin with the LHC, I must wait for some new, hopefully raw, data from it. Can I come to a conclusion from it? I don't know...

Random generated using unicellular organisms also qualify for a good, practical random but every life form including the unicellular organisms are controlled by their genetic makeup and this determines their life, including movement. In case of a unicellular organism, they are in fact not free to float, but move towards chemical or physical gradients, where chances of survival are more. But still, it creates a good random...

www.random.org creates random numbers using the physical variable of atmospheric noise that can be picked up from a normal radio or perhaps with any good quality microphone. If we use this as a seed and do hashing on it, we will get a relatively higher degree of randomness.