Profilo di ChaosConstLatest news from ChaosCo...BlogElenchi Strumenti Guida

Blog


20 novembre

Some Ideas For EE PartII/III...

 
1.x[n]=f(x[n-1])的历史
 
如果我们想从抽象的数据和规则入手,感觉就像寻找一个初始向量X[0],一个规则f.
然后系统就开始运转,x[n]=f(x[n-1]).
 
我不太清楚CPU的运转方式,但我仍以为可以把内存和硬盘包括寄存器等等的全部数据看作一个初始向量X[0],那么规则f就是指令系统等计算机运转的法则,X从X[0]开始,在f作用下的的第n次改变后的数据集合记为X[n],于是一个计算机运转的过程就可以看作一个向量函数迭代.x[n]=f(x[n-1]).
把世界上所有计算机看作向量X,把计算机运转规则和Internet规则(IEEE802.x之类)看作f,那么整个Internet的运转也可以看作向量函数迭代.x[n]=f(x[n-1]).
这样就完了吗?不,把整个宇宙的基本结构(夸克一类)的全部集合(如果有的话)的一种状态(就算一种可能的状态也行)看作向量X,把一统一理论(从相对论量子力学到十一维空间,如果有的话)看作法则f,那么这个世界的运转也可以看作向量函数迭代.x[n]=f(x[n-1]).(你看到这里会不会有不屑的神情?但我有想法就要说出来,我不怕出错.)
 
我认为任何一个过程,都可以抽象的表达为x[n]=f(x[n-1]).(当然,f不是普通的向量函数,至少要具备随机性).
任何X,f,都可以看作某种过程的表达.
 
一般的过程不是我们所关心的,我们只关心能产生生命的过程,我们相信产生生命是产生智能的前提.
世界的运转被证明是可以产生生命以至智慧的,但它是上帝创造的.
我们相信Internet的运转这个过程已经足够大也能够产生生命甚至高级的智慧形式,我们已有f,需要X(有模型的网络空间).一个充斥着具有生命的可繁衍的程序的Internet.
为了研究X,我们试图缩减f的规模,试图找到适合单个计算机运算法则的X,只要X有生命.
但是这样仍然太难,我还试图进一布缩小f的规模,寻找更小规模但又不脱离计算机运算法则的f,寻找合适的X,寻找合适的向量函数迭代.x[n]=f(x[n-1]).
 
2.平坦数据和稳定进化
如果真的向量函数迭代就是我们的目标的话,那么必须明确的知道怎么样的f,怎么样的X是我们所需要的.
用话来说就是:
1)规则f是不变的
2)规则f是与CPU指令系统基本相适应的
3)X中没有程序与数据的区别(平坦数据)
4)X逐渐趋于有序(稳定进化)
5)X中形成了独立于f的可以遗传变异的规则(有生命)
我不能说满足以上5条就一定成功,但是我肯定没有以上5条一定不成功.
 
以下才是重点:
什么叫规则f是不变的?
对于相同的一个向量X,无论在怎样的情况下用f进行作用,都能得到相同的结果分布.
用f对X作用,可能得到X1,X2,X3,其中20%的可能得到X1,30%的可能得到X2,50%的可能得到X3.这就是一个结果分布.
 
现在说核心内容:
什么叫没有程序与数据的区别?
用话来说:
1)X有无限变化的可能
2)X可以分割
解释一下:
1)
X的所有可能取值构成一个集合V,对于单个计算机来说,就像内存和硬盘中所有可能存入的内容.
设V(i),V(j)为V中任意两个元素,再设X[0]=V(i),X[n]为向量函数f以X[0]为初值迭代n次后的结果.
存在N足够大,当n>N时,P{X[n]=V(j)}>0 |(X[n]取V(j)的可能性大于0)
以单个计算机为例,上面那段话的意思是无论我们在硬盘和内存中写入什么数据,我们总可以把它变成另外一种指定的数据.(可能不是这样,那么单个计算机的内存和硬盘在以指令系统为规则下就构不成平坦数据)
2)
严格的数学定义(略)
以我们的世界为例,就是说,你把世界分割以后,它还是一个世界.这个被分割的世界会受到分割的影响,但是这种影响与分割的大小成反比,与分割的位置无关.
反观传统的程序与数据分离的模式,例如坦克大战,环境数据放在前面,坦克数据放在后面.如果做分割,割在后面的数据永远的不到环境数据,永远无法正常运行.
虽然我希望这就是定义,但很遗憾不是,以上两条我以为可以构成平坦数据的必要条件.

什么叫逐渐趋于有序?
把X[n]=f(X[n-1])的一次迭代看作一次试验,把从f(X[n-1])的可能结果看作试验的结果.引用香农信息论中的熵的定义,称H(n)为得到X[n]的熵.
则H(n)有逐渐减小的趋势.
21 giugno

A general outline of my first thesis of EEC

//June 11, 2006

A general outline of my first thesis of EEC

 

Evolution of Electronic Creature (How it works, and how it will work.)

 

1.    Introduction: Hi! Want to create your own world?

2.    Feasibility: Simulation of nature & computer. (How does our nature & computer work?)

3.    Goal: Out of your imagination.

4.    Summary of Part I: First thing’s first

5.    Overall of Data structure & Arithmetic frame. (Hard work: What is we need?)

6.    Prisoners’ dilemma: Fish’s funny puzzle. (Game theory & Nash equilibrium)

7.    Codon List and CPU repertoire. (Write your own RNA!)

8.    Fish’s data. (Your Adam.)

9.    Ev data & regulars. (Let there be light.)

10.   Heredity & aberrance. (Evolution comes from here)

11.   EE Program: We are go for launch!

12.   Experiments: Let’s breed!

13.   Conclusions. (See our achievements)

14.   Philosophy: (God: ourselves?)

15.   Expectation: (I have a dream.)

 

我的第一篇電子生物論文之大綱[繁體中文版]

 

電子生物進化方案(如何運轉以及將如何運轉)

 

1.     引言:你好!想要創造一個你自己的世界?

2.     可行性:自然界與電腦的相似性(我們的自然和電腦如何工作?)

3.     目標:超乎想象

4.     第一部分概說:萬丈高樓平地起

5.     資料結構與演算法框架總攬(艱難的工作:我們需要什麽?)

6.     囚徒困境:魚類有趣的困惑(博弈論與納什均衡)

7.     密碼子表與指令系統(編寫你自己的RNA!)

8.     魚的資料(女媧造人)

9.     環境資料與規則(盤穀開天)

10.   遺傳與變異(進化自爾起)

11.   電子生物程式:只欠東風!

12.   實驗:開始培育!

13.   結論:(看看我們的成就)

14.   哲學:(上帝:我們自己?)

15.   展望:吾將上下而求索

 

28 marzo

Study Record

How to realize EEC (Part 2)

After about 10 hour’s thinking, I got this:

 

 

Static Data

Dynamic Data

RNA of Creature

(SR1) Heredity regulars

(DR1) data in register

(SR2) List of Codon

(DR2) RNA codes

Environment

(SE1) Ev constant

(DE1) Ev Dynamic Data

(SE2) Ev regulars

(DE2) Creature’s Ev Data

 

 

Before we continue, I want to emphasize one thing: we must build a model, which can adopt many “world” (as many as possible). And the model must simple enough, or we can’t make it.(as simple as possible).  Universal <==>Simple

 

See a graph:

 

It shows the relations between “~or”(SR2 Editor, RNA&DE Editor, RNA translator and Environment Processor) and data (SR2, SE, DE and DR). Each line means the information flow from arrowfoot to arrowhead.

It’s really beautiful, is it?

 

 

Arithmetic frame:

 

unit SR2;

type

  //structure of DR1

  ...

  //structure of DR2(RNA)

  ...

var

  ...

begin

  ...

  Case DR2 of

  000000(AAA):

        begin

          ...

        end;

  000001(AAU):

        begin

          ...

        end;

  ...

  end;

end.

 

unit Environment Processor;

 

function SR1;

 

program Main (Output);

uses SR2 (List of Codon) , Environment Processor;

type

  creature = record

               RNA:array[1..10000] of byte; //DR2

               ...                          //DE2 

             end;

  things1 = record

              ...

            end;

  ...

 

var

  creatures : array[1..10000] of creature;

  thing1s   : array[1..100] of thing1;

  ...

 

const

 

  SE1 = ...

  ...

 

begin

  //Initialize data

  RNA&DE Editor;

  repeat

    repeat

      Environment Processor;

{

      1. Process every rules in the world (only in one unit time).

2. Process each Creature's action.

          If the action is “breed”, uses SR1 function;

}

      RNA translator;

{

       For each creature

       begin

          put DR1 in memory;

          for Num_process (How many nucleotides translate In one unit time, decided by SR2)

          begin

            according to SR2 & RNA, translate RNA->Pascal Code, and run it.

          end;

          memory write into DR1

       end;

}

      DE viewer (inside RNA&DE Editor);

    until PAUSE;

    RNA&DE Editor;

  until EXIT;

end.

 

 

NOTE: The frame shows how whole system works. Inspire we only have one thread in it, we simulate multi-thread running. Actually, not only RNA translator and Environment Processor work independently. But also every creature translates their RNA-Pr and every thing in the world move at one time! So multi-thread is very important, But I just have one computer, which only has one CPU, what can I do? Is there something different between use multi-thread programs in my computer and simulate multi-thread?

… May be I should use network in Part III?

 

11 marzo

Evolution of Electronic Creature (Study Record, and sth. like that…)

Evolution of Electronic Creature (Study Record, and sth. like that…)

 

Goal:

Part 0

0.     EE virus.(TOP SECRET)

Part 1

1.        Make a program to find a batter strategy in a certain circumstance.

2.        Make a creature exchange message with environment program.

Par 2

3.        Make a digital world all in it is 01010011~, but under Some Rules, real creatures living in.

 

Summary of Two Parts:

 

1.        Build creatures can exchange message with environment program; this creature has its own protein code that decided its all habit. It has three characters followed:

(1). Reaction rapidly with changes of environment.

(2). Adaptation.

(3). Heredity and aberrance.

This part of program has its structure of datas.

(D1). Data of creatures’ Pr.

(D2). Rule of how Pr works. (Just like CPU repertoire)

(D3). Data of environment.

In this part I will make three independent programs:

(1). CreatureMind: Give action decided by D1 to Environment.

(2). Environment: Receive and process action from CreatureMind and offer environment messages to CreatureMind.

(3). Pr maker: Use to Create a Creature’s Pr more easily.

 

2.        Build a digital world and a Rule that the world is running under. Creatures are made of 0101101~, and environments is also made of these numbers. In this part, creatures and environment is made of the same thing! And they are running under the same rule. So the structure of datas is reduce to this:

(D1). Data of creatures & environment.

(D2). Rule of how D1 works. (Also like CPU repertoire)

As creature and environment becomes the same thing, creatures’ structure became freer and complex; and can have real biological creature properties as follows:

(1). Reaction rapidly with changes of environment.

(2). Adaptation.

(3). Information flowing. (Metabolism)

(4). Same basic structure like cells.

(5). Heredity and aberrance.

(6). DNA->Pr

In this part I will make only two independent programs:

(1). Creature &Environment: Change world under D2.

(2). DNA maker: Use to Create a Creature’s DNA more easily.

 

NOTE: As you see, Part2 is expansion of Part1. And Part1 is much easier than Part2, and I am very confident about Part1, I will prove it’s right. But Part2 is much exciting than Part1, even it’s very hard to archive, I will keep running.