15 May 2010, 7:19 pm
I actually like the idea as I've been thinking about something like this before. I'm very interested in AI, especially in the implementation of human-like logic and intelligence (evolutonary intelligence).
It's just that I'm probably not quite experienced enough to do real complex stuff like this. Oh, and I'm also just too busy at the moment - so if I were to attempt something that would be in the summer break 
With the brainstorming I could help, though - as I know how things work concerning programming.
Not now, anyway... 
(It's 2:15 Sunday morning by now)
15 May 2010, 7:58 pm
16 May 2010, 4:05 am
Ah, well
What you are doing now, is questioning the definitions 
Indeed we cannot really say what intelligence is, because there are so many things to measure it by, but that's only valid for intelligence tests for, say, IQ and EQ
Intelligence how I meant it here, is the working of the outter parts of our brains; the parts that enable us to think abstractly and give us an advantage over animals. The working of this is already known quite well, only it is hard to put it into practice:
The human brain works like an evolutionary machine; young children know nothing, and work with the process of trial and error. With every impulse they receive, some signals are - quite randomly - sent through the brains, and only the ones that have a positive effect (in fulfilling their basic needs) return a signal to strengthen the connection between these two neurons.
I do not know the exact story, though.
If you want to know more, google neurobiology 
Plus, I am interested in AI, but that does not directly mean the only thing I want to do, is build programs with it. In fact, this is way too difficult for me to do, so I usually stick with very simple simulations. But I am more interested in the research of the brain, in the underlying science. And I do not actually believe anything rivaling human intelligence could be made within the first few decades, as the human brain has a 'processing time' way above that of computers. That should not stop us from applying it in a simpler form for useful purposes, though...
But back on topic;
I imagine the model as some sort of associative database; That is, a database not indexed and accessed as if being a list of items, but rather as if being a network of nodes which are connected with each other. Besides that every connection can have some data, just like every node.
Or say that you prefer it mathematically; (I made up my own notation, though!)
N is a node in the network, just like P
N itself has a value, which for the purpose of explaining this does not matter, but which would represent something like an identity.
Idem for P.
[NP] is the connection between N and P, which also has a value, representing the strength of the relating, which is probably related to how close their individual values lie from each other
Problem is, though; there should be the connection for it to grow/decay, so before you even set up rules for this, you need to know how connections appear. This is not quite easily simplified, but there must be some way to represent this behaviour approximately, as in theory it is not that hard:
1. People need to be physically able to reach each other (Time, Place, state of mind, physical state, etc. )
2. Circumstance should allow a conversation to take place
3. There should be no outside influences preventing the conversation from proceeding naturally
4. It should be possible to stay in contact with the other, afterwards
I'd like to know what you think of it...
16 May 2010, 6:19 am
16 May 2010, 6:24 am
16 May 2010, 6:26 am
16 May 2010, 6:58 am
You're the first person in this thread who's actually given indication that they understand what I'm talking about. You're also the only person to add anything constructive beyond "OMG STOOPID WON'T WORK! !1".
People in this thread don't seem to realise that models are inherently imperfect because they are not the things they are modelling. But if they were the things they were modelling they they wouldn't be models! The key is to capture the pertinant aspects and "pertinant" is dependant on context and, I suppose to some degree, subjective.
The problem is that I have actually studied these kinds of things. These kinds of things are being attempted by research mathematicians (look at say systems biology) and are completely beyond your reach. Modelling society is a complete waste of time as things stand, or indeed relationships.
You should realise that to play the social game and win, you would need a near perfect model to do this. Something which you have no chance whatsoever of making.
I assume the use of the model is not determining a specific situation, but rather studying how certain social processes work in general. If I were to make a program that simulated evolution, I would by no means do this hoping to figure out how mankind would look in about 10000 years. I would rather do so to take a closer look at the workings, as looking at a model actually teaches us a lot about the system. Moreover it enables simple experiments to be done, without the need of thousands of participants.
In other words: it is rather meant as a sandbox program intended for research, than as a huge calculator predicting our social future.
And still, even if it is a complete waste of time, it is of no use to keep telling that. If you think it is a waste, just don't react.
The request of the TS was for people's ideas, not their disapprovement
16 May 2010, 7:50 am
16 May 2010, 7:52 am
16 May 2010, 10:55 am
You're the first person in this thread who's actually given indication that they understand what I'm talking about. You're also the only person to add anything constructive beyond "OMG STOOPID WON'T WORK! !1".
People in this thread don't seem to realise that models are inherently imperfect because they are not the things they are modelling. But if they were the things they were modelling they they wouldn't be models! The key is to capture the pertinant aspects and "pertinant" is dependant on context and, I suppose to some degree, subjective.
The problem is that I have actually studied these kinds of things. These kinds of things are being attempted by research mathematicians (look at say systems biology) and are completely beyond your reach. Modelling society is a complete waste of time as things stand, or indeed relationships.
You should realise that to play the social game and win, you would need a near perfect model to do this. Something which you have no chance whatsoever of making.
I assume the use of the model is not determining a specific situation, but rather studying how certain social processes work in general. If I were to make a program that simulated evolution, I would by no means do this hoping to figure out how mankind would look in about 10000 years. I would rather do so to take a closer look at the workings, as looking at a model actually teaches us a lot about the system. Moreover it enables simple experiments to be done, without the need of thousands of participants.
In other words: it is rather meant as a sandbox program intended for research, than as a huge calculator predicting our social future.
And still, even if it is a complete waste of time, it is of no use to keep telling that. If you think it is a waste, just don't react.
The request of the TS was for people's ideas, not their disapprovement
I wasn't thinking along the lines of an analytic solution, or even a numerical one. Rather, I was thinking about general behaviour, which is hard enough to do even with a model of three equations.
But if anyone did want to sandbox, here is what you do. Make an account on an internet forum full of NT's somewhere (or even WOW or Second Life would serve a similar function - indeed some researchers have been looking at doing just this to avoid having to model). Say some stuff. See how they react. When they get fed up, make another account and you are back in the game. Far more effective than any model.
16 May 2010, 11:02 am
17 May 2010, 7:32 am
17 May 2010, 9:22 am
17 May 2010, 10:23 am
17 May 2010, 11:11 am
