Introduction to my vision of AI and 
Learner, which is the first attempt to build a practical application
, in the purpose of evaluating ideas.

PART 1

Definitions and functionality:

AB  	 Artificial Being. 
     	The AB is the agent which operates in the world and
     	is characterized by the following

	1) fixed senses and motorics (input / output)
	2) innate(fixed) knowledge at basic levels 
	3) knowledge generated by interaction with the world, experience.

	Keeping a clear line between innate (fixed) and 
	experienced (dynamic) knowledge is in my opinion of greatest importance.

	4) The pattern detector generates knowledge/rules from repetetive
	   patterns and is to be considered innate at lower abstraction levels.
	   By lower abstraction levels i mean near the senses

	5) A stimulator, which uses random output in situation where
           no knowledge exists and in the initial phase of the AB

WORLD  Is the place where our AB exists.
	It can be the real world or a virtual one.
	Learner consists of a very simple virtual world and an AB trying
	to understand its surroundings and itself.
	The world is characterized by natural laws,that the AB cannot override. 

 This picture illustrates 
 the world.
 The AB is located at
 X=1, Y=1

 The AB has 1 sense,
 it senses the value at its
 current position = 5

    	The 'Motorics' are     1 = UP   2 = DOWN  3 = LEFT  4 = RIGHT
	The only natural law that exists in the world is world boundaries
	iow, the AB cannot escape the world.

PRIMARY DATASTREAM

	When the AB moves a primary datastream is generated from motorics and senses and 'time'
	time in this case is the datastream index

	Example:

       Time   Output  Input
	0		5      	(initial condition)
	1	1	2
	2	4	3
	3	4	3	(cannot escape world)
	4	2	6
		etc

CONSIDERATIONS
	
	1) How can we extract knowledge from this stream ?
	2) How shall we represent the knowledge
	3) What innate knowledge should we have ?

INNATE
	1) input / output capability (primary datastream)
	2) > , < , = , <> , AND
	3) algorithm that generates primary knowledge

REPRESENTATION OF INPUT / OUTPUT AND TIME

	S(t) represents the input value at time(step) t
	M(t) represents the output value at time t
	S(t):x represents the fixed input value x at time t
	M(t):x represents the fixed input value x at time t

REPRESENTATION OF PRIMARY KNOWLEDGE

	example 1
		S(0):5 AND M(1):1 AND S(1):2    
		time is relative to any given position
		When translated to english this knowledge (rule)
		says: When sensing 5 and moving up, I will sense 2

	example 2
		S(0):9 AND M(1)<S(1)    
		time is relative to any given position
		When translated to english this knowledge (rule)
		says: When sensing 9 any motoric (action) value
 		will be less than the sensed value






Back to Main