Geth programming language
#1
Posté 26 février 2010 - 04:39
If we have some geeks in here I need some representation of how such a Geth programming language would look like. Not actual "real" programming language.... Just for fun
Just how your interpretation of such a language between 1,183 active AI programs would look like.
It needs to be understandable to people like me.. Who isn't all the techy... Remember I need it for role-play purposes. For fun.
I was thinking about making some modifications to a HTML code, so it looks like Geth are gaining consensus.. Talking to eachother... And then post it in my character biography.
But I don't really know how....
#2
Posté 26 février 2010 - 04:43
Could use Prolog (wikipedia should have some examples) as a good (read: fairly easy to understand) representation of an AI's knowledge representation and logical processing systems.
Modifié par Domcor, 26 février 2010 - 04:44 .
#3
Posté 26 février 2010 - 04:46
Each Geth program would be an artificial neural network, and the different networks together form a meta-network.
To humans, the networks would appear to be nonsensical black boxes, since they are arranged in the most efficient way rather than the most understandable (for humans) way. Also, they'd have the ability to evolve and adapt.
#4
Posté 26 février 2010 - 04:53
I Just need a graphical representation of how it would look like. In text.
#5
Posté 26 février 2010 - 04:54
Windows are structural weaknesses. Geth do not use them.
#6
Posté 26 février 2010 - 04:56
Tensdale wrote...
You're getting waaaay into it. It doesn't really matter if it doesn't fit up to any "truth" standard about AI intelligence.
I Just need a graphical representation of how it would look like. In text.
01010100 01100101 01101110 01110011 01100100 01100001 01101100 01100101 00100000 01110111 01110010 01101111 01110100 01100101 00101110 00101110 00101110 00001101 00001010 00001101 00001010 01011001 01101111 01110101 00100111 01110010 01100101 00100000 01100111 01100101 01110100 01110100 01101001 01101110 01100111 00100000 01110111 01100001 01100001 01100001 01100001 01111001 00100000 01101001 01101110 01110100 01101111 00100000 01101001 01110100 00101110 00100000 01001001 01110100 00100000 01100100 01101111 01100101 01110011 01101110 00100111 01110100 00100000 01110010 01100101 01100001 01101100 01101100 01111001 00100000 01101101 01100001 01110100 01110100 01100101 01110010 00100000 01101001 01100110 00100000 01101001 01110100 00100000 01100100 01101111 01100101 01110011 01101110 00100111 01110100 00100000 01100110 01101001 01110100 00100000 01110101 01110000 00100000 01110100 01101111 00100000 01100001 01101110 01111001 00100000 00100010 01110100 01110010 01110101 01110100 01101000 00100010 00100000 01110011 01110100 01100001 01101110 01100100 01100001 01110010 01100100 00100000 01100001 01100010 01101111 01110101 01110100 00100000 01000001 01001001 00100000 01101001 01101110 01110100 01100101 01101100 01101100 01101001 01100111 01100101 01101110 01100011 01100101 00101110 00001101 00001010 00001101 00001010 01001001 00100000 01001010 01110101 01110011 01110100 00100000 01101110 01100101 01100101 01100100 00100000 01100001 00100000 01100111 01110010 01100001 01110000 01101000 01101001 01100011 01100001 01101100 00100000 01110010 01100101 01110000 01110010 01100101 01110011 01100101 01101110 01110100 01100001 01110100 01101001 01101111 01101110 00100000 01101111 01100110 00100000 01101000 01101111 01110111 00100000 01101001 01110100 00100000 01110111 01101111 01110101 01101100 01100100 00100000 01101100 01101111 01101111 01101011 00100000 01101100 01101001 01101011 01100101 00101110 00100000 01001001 01101110 00100000 01110100 01100101 01111000 01110100 00101110
#7
Posté 26 février 2010 - 05:07
eg.
Consensus* consensus = new Consensus("Some sort of problem");
consensus->Evaluate();
consensus->Query();
return consensus->Output()
Short of some sort of binary representation (or really complex AI algorithms) youre not really gonna find particularly good representations.
ps. From an RP standpoint; whilst most appropriate, I find binary in AI players' bios to be really tacky.
Modifié par Domcor, 26 février 2010 - 05:08 .
#8
Posté 26 février 2010 - 05:08
100100110010100100000011001100111010101101110011011100111100100101110000011010000101000001
10100001010010100110110010101110010011010010110111101110101011100110110110001111001001011
1000100000010001110110111100100000011001100111010101100011011010110010000001111001011011
1101110101011100100111001101100101011011000110011000101110
Modifié par Tensdale, 26 février 2010 - 05:08 .
#9
Posté 26 février 2010 - 05:13
110100101110010011001010110001101110100001000000100001101101111011011100111010001
1100100110111101101100
Modifié par rabbitchannel, 26 février 2010 - 05:14 .
#10
Posté 26 février 2010 - 05:19
The question is, do you really need imaginary lines of code that represent the AI programs or do you actually want to script a conversation between them?
#11
Posté 26 février 2010 - 05:30
#12
Posté 26 février 2010 - 05:33
Domcor wrote...
Well, if you don't want something realistic, then some sort of abstract object oriented approach could work.
eg.
Consensus* consensus = new Consensus("Some sort of problem");
consensus->Evaluate();
consensus->Query();
return consensus->Output()
Short of some sort of binary representation (or really complex AI algorithms) youre not really gonna find particularly good representations.
ps. From an RP standpoint; whilst most appropriate, I find binary in AI players' bios to be really tacky.
I like this. But I would need ALOT more about this to actually twist it together with a certain situation. And I have no idea what code you got the idea from? html?
#13
Posté 26 février 2010 - 05:43
basically creates a "Consensus" object and tells it what the problem is.
calls the Evaluate method (which i imagine probably asks each individual geth to determine what the solution to the problem is).
calls the Query method (which compiles the individual solutions into a single solution).
returns the solution.
It probably has no real relation to what would actually happen in an AI system (I've only done Intro to AI on my compsci course and am gonna start actually programming for it next year), but seems abstract enough that it would work for RP purposes.
#14
Posté 26 février 2010 - 05:46
#15
Posté 26 février 2010 - 05:56
INPUT:: "Tensdale, where should we go? Point A or Point B?"
Consensus* consensus = new Consensus("Optimise <Point A> <Point B>");
consensus->Evaluate();
consensus->Query();
return consensus->Output()
OUTPUT:: "Optimal <Point B>"
#16
Posté 26 février 2010 - 08:36
As said before I just need imaginary lines of code that represent the AI programs.. But it should "look" like computer codes.. Remember this is just for a game. For fun.. The code doesn't have to technically mean something. But look like it does, and it should also be understandable to everyone what exactly it means
#17
Posté 26 février 2010 - 10:48
Tensdale wrote...
Seems a little hard to understand for people who don't understand c++..
As said before I just need imaginary lines of code that represent the AI programs.. But it should "look" like computer codes.. Remember this is just for a game. For fun.. The code doesn't have to technically mean something. But look like it does, and it should also be understandable to everyone what exactly it means
So you are asking for a code that has almost 2000 processes, it should look like real code, but not be real code and be understandable to people who don't know jack about programming.
Do you see the problem here?
Just wing it. A geth code would not look like anything programmed now or programmed by humans. Just write text that looks like computer code to you, that's the only way you would understand it. None of the current programming languages are easy to read for someone who has no idea how they work.
So just write some crap with some ( ) and [ ] and { } and end line you write with ; and your "code" will look enough like real code. And most of all you can tell you friends what it's supposed to do.
#18
Posté 26 février 2010 - 11:16
Evaluating Data...
Querying Geth Nodes...
Building Consensus...
Consensus Reached...
Compiling Solution...
Solution: Point B is optimal
#19
Posté 27 février 2010 - 01:06
6_|-/\\-/.U|(X)UU|.??(X)u-|-uu|.-|-(X)(X)>|.-|-/>|_|-./\\-//nu+
U|<#>(|)./.nu-|-_|-_|-4.??//\\-/.(|Y|)<#><|<>-|-_|-nu+
Data packets are sent to communicate with the neural network, they have source/destination identifiers, instructions about the packet's contents, error control info (checksums) and then the bulk data that gets processed on the receiving end. An example is if you are updating the neural network with a piece of info that says to avoid certain planets, you send it out to all destinations (all geth) and they all use the data, but if one geth is in a particular place and data is only relevant to him then you have it flag only him as the destination and the rest of the geth ignore the packet.
I know this. My sister was a geth.
#20
Posté 27 février 2010 - 01:15
Tensdale wrote...
With the example you've created here, how would it look like if Legion should decide to go to pint A or to point B, where the majority of the 1,183 active programs have decidedd that point B is the best path?
Easy, in that case Legion would straight away go for point B. As a synthetic he is unable to get drunk and has no use for pint A. Sorry, couldn't resist...
#21
Posté 27 février 2010 - 01:34
Also you're probably right guys. I'm probably trying to do this too simplified.. I'll see if I figure out a solution. Thanks for all the input. It's gonna be usefull.
#22
Posté 27 février 2010 - 03:10
#23
Posté 27 février 2010 - 06:21
Recnamoken wrote...
Executing sudo command.
Windows are structural weaknesses. Geth do not use them.
hehe that includes windows vista and 7
#24
Posté 01 mars 2010 - 01:42
The example in the article should be easy to understand, even with no experience in programming.
I wouldn't use html or xml, those are mainly used for displaying information, while scripting/programming is used when the computers "think".
#25
Posté 01 mars 2010 - 02:01
If this is so, odd syntax and a hurry-up style a-la Mordin would be most appropriate.
Perhaps something like this for exploring a life bearing world with only plant analogs.
"<mountain green> <rain heavy> <null life animal> <life phytophillic> <unusual significant>"




Ce sujet est fermé
Retour en haut






