PDA

View Full Version : Neuronal Nets In Objects (PhysMathChemMedicChaos...)



Zander
01-05-1995, 08:23 PM
----------------------------------------------------------
! H E L P !!! !
! !
! Perhaps many new scientific discoveries are possible!! !
! Please send this text to scientists !
! (physicists, chemists, medics, engineers ...), !
! scientific instituts, firms etc. or make this !
! text available. The text is N O T only for !
! specialists! !
! This theory is not much known. Nobody knows this !
! simple theory. !
! !
----------------------------------------------------------
(I'm new in Internet)

Please send your opinion to news: sci.chaos
Use the word NEURO-OBJ in subject!

Can you create a new newsgroup: sci.chaos.neuro-objects
If this newsgroup exists then send your opinion to
this newsgroup.

E-Mail adress of the writer: zander@rz.uni-leipzig.de

================================================== =============
Neuronal Nets In Objects
================================================== =============
It's only a theory.

The following suggestions are not new. But they are not
much known. Ilja Prigogine (Nobel Prize Winner) and other
people said simular suggestions.
Please go ahead with own reflections and do experiments
and computer-simulations. Perhaps many new scientific
discoveries are possible!


----------------------------------------------------------------
Neuronal Nets In Mixtures Of Substances
----------------------------------------------------------------

Perhaps you can produce Neuronal Nets in mixtures of substances
with relatively changeable properties and substances with
memory properties.

X O X O X O
O X O X O X
X O X O X O
O X O X O X

X - Particle with relatively changeable properties
(A cell of the Neuronal Net)
- Its state fluctuates.

O - Particle with memory properties
(This particle cause that the kind of connection
between cells "X" is relatively constant.)
- According to memory state of this particle the
state of a cell determines the state of the
neighbouring cell.
- High-energy (chaotic) successions of cell states
can change the memory state of this particle.
Because of that evolution by energy control is
possible.
- Somtimes Rule of Hebb exist:
The same state of neighbouring cells leads to
another memory state of this particle then
different states.

For example:
-----------
- A mixture of a conductive substance (eletric current is
changeable = "X"-particles) with a magnetic
substance (memory properties = "O"-particles)
- Connection beetween cells "X" takes place by induction
(according to memory state (polarization) of the
neighbouring particles "O").
- Evolution by energy control: Energy supply takes place
by a global alternating field. High-energy supply leads to
chaotic currents --> New Order is possible. Low-energie
supply --> strengthens the old Order.
(read the chapter "The Learning Process")
- Rule of Hebb: The same current direction of neighbouring
cells "X" leads to another polarization of the
neighbouring particle "O" then different current direction.


----------------------------------------------------------------
Neuronal Nets In Any Substances
----------------------------------------------------------------

Perhaps you can produce Neuronal Nets in any substances,
too, because molecules have relatively changeable and
relatively constant (memory) properties.

Distribution of properties "X" and "O" in a structure
of several molecules:

X O X O X O
O X O X O X
X O X O X O
O X O X O X

X - relatively changeable property of a molecule
(A cell of the Neuronal Net)
- Its state fluctuates.

O - memory property of a molecule
(cause that the kind of connection between cells "X" is
relatively constant.)
- According to memory state of this "O" the
state of a cell determines the state of the
neighbouring cell (physical interaction).
...
... (Analogie to "Neuronal Nets in Mixtures")

----------------------------------------------------------------
The Learning Process
----------------------------------------------------------------

The learning process is stimulated by global energy supply.
It's simular to brain: If something is not OK then energy
supply is increased. If everthing is OK then energy supply
is decreased.

---------------------
! global energy supply!------------
------------- feedback


A T F I R S T there are only random fluctuations. Each system
fluctuates in a somewhat chaotic way:

1.) Sometimes the system fluctuates in a undesired direction:
--> Energy supply is increased
--> The neuronal connections are destroyed partly
(Not all connections are destroyed: only "bad",
unstable configurations of connections)

2.) Sometimes the system fluctuates in a desired direction:
--> Energy supply is decreased
--> The neuronal connections are strengthened partly


1. + 2.) ---->>> It's simular to selection process
in nature (E V O L U T I O N)!!!
--------------------------------

L A T E R : desired behaviour is strenghened.


Remark:
------
- You have to find the middle energy supply and the variation
of energy supply by measuring series or by flexible
computer programs.



----------------------------------------------------------------
Teach The Object Any Physical Or Chemical Properties
----------------------------------------------------------------
Perhaps you can teach the object any physical or chemical
properties?
It's N O T necessary to know the following
mechanism, because self-organization takes place.

At First :
----------
Chance distribution of any L O C A L properties "a" and "b"
in the object:

Object: (It's only an example)

abababbabbaba
bababbababbab
babaaabbabbaa
bbabbbabbbbab


1. Possibility (After learning process):
----------------------------------------
Moving balance of L O C A L properties (It's a very simple
learning destination)

bbbbbabbaaaaa
bbbbbbabbaaaa
bbbbbaababaaa
bbbbbbaaaaaaa

--> other G L O B A L physical and chemical properties


2. Possibility (After learning process):
----------------------------------------
Moving space and time concentration of L O C A L proterties

bbbbbaaaaabbb
bbbbbaaaaabbb
aaaaabbbbbaaa
aaaaabbbbbaaa

--> other G L O B A L physical and chemical properties

3. Possibility:
--------------
Perhaps you can control known self-organization-mechanism
(for example Belousov-Zhabotinsky-reaction) to disired
direction by learning process.

Remark
------
It could be that "a" and "b" are unknown physical properties in
microcosm. Then global properties (after learning process) are
unknown, too. ---> New scientific discoveries are possible!!



----------------------------------------------------------------
Literature
----------------------------------------------------------------

Read literature about Chaos Theory, Neuronal Nets (especially
Hopfield Net) and evolution!


----------------------------------------------------------------
Possible Uses
----------------------------------------------------------------

1.) Information Processing

---------------------
! global energy supply!------------
------------------- feedback


2.) Perhaps you can teach the object any physical
or chemical properties. For example: improvement
of energy transformations, for example a photo element:

------------------------
! global energy supply !--------------
------------- feedback


3.) improvement of healing processes

------------------------
! global energy supply !--------------
! instruments ! feedback
-------------

(Perhaps in some cases you can control the energy supply
by mental concentration (without any machines)?)


4.) You have to apply this theory to your own special problems.
(If the measuring instruments record an approach to desired
behaviour then energy supply is decreased. Otherwise
energy supply is increased.)

5.) Perhaps you can teach the object unexpected, remarkable or
"impossible" properties.


----------------------------------------------------------------
An Experiment For You: A Chaos-Game
----------------------------------------------------------------

At first the game figures are distributed by chance.

Destination of the game:
The number of the figures on the left side of the
game board should be greater then number on the right side:
number LEFT > number RIGHT (moving balance)

------------------------
! global energy supply !
! (You shake the game !
! board with your hand.)!--------------
! instrument ! feedback
! (your eye) !
-------------


The experiment shows:
Evolution by energy control is possible.

You have to imagine:
a game figure "O" = local memory property "O"
in this game in a substance


------------------------------------------------------------------
Computer Simulation Of This Chaos-Game
------------------------------------------------------------------

uses crt;
const maxcell = 10;
var board:array[1..maxcell,1..maxcell] of integer;
sum_left,sum_right,number:real;

procedure chance_direction (var dx,dy:integer);
(*-------------------------------------------*)
begin
dx := - 1+trunc(random (300)/100);
dy := - 1+trunc(random (300)/100);
end;

procedure chance_place(var x,y:byte);
(*---------------------------------*)
begin
x:= random (maxcell) + 1;
y:= random (maxcell) + 1;
end;

procedure init;
(*------------*)
var x,y:byte;
begin
sum_left := 0; sum_right := 0; number:=0;
randomize;
for x := 1 to maxcell do
for y := 1 to maxcell do
board[x,y] := random(2);
end;

procedure drawing;
(*--------------*)
var x,y:byte;
begin
randomize;
for x := 1 to maxcell do
for y := 1 to maxcell do
begin
gotoxy(x*4,y*2); write (board[x,y]:3);
end;
end;

function strength_of_energy: integer;
(*----------------------------------*)
var number_left, number_right:integer; x,y:byte;
begin
number_left := 0;
for x:= 1 to maxcell div 2 do
for y := 1 to maxcell do
number_left := number_left + board[x,y];

number_right := 0;
for x:= maxcell div 2 + 1 to maxcell do
for y := 1 to maxcell do
number_right := number_right + board[x,y];

strength_of_energy :=
round(0.25* (number_right-number_left) + 10);

number := number + 1;
sum_left := sum_left + number_left;
sum_right := sum_right + number_right;
gotoxy(65,1); write(number_left:4,' ',number_right:4);
gotoxy(65,3); write((sum_left/number):4:1,' ',
(sum_right/number):4:1)
end;

procedure shake (strength:integer);
(*-------------------------------*)
var i,dx,dy:integer; x,y:byte; new_x,new_y:integer;
begin
if strength 0 then
begin
dec(board[x,y]);
inc(board[new_x,new_y])
end;
end;
end;
end;

(* main *)
(*------*)
begin
clrscr;
init;
repeat;
drawing;
shake(strength_of_energy)
until keypressed;
end.




writer: Carsten Zander, Germany