THE  ' SAI  CORE '  DESIGN

   GENERAL ARCHITECTURE:
	OBJECTS WITH ASSOCIATIONS  AS  GRAPH/TREE

   ENGINE:
	GOAL ACHIEVING AND FORMING
	PERSISTENT BASIC GOALS
	KNOWLEGE PRESENTATION THROUGH MODELS &RELATIONS
	KNOWLEGE  ACQUISITION,  SEARCH,  MODEL APPLICATION
	KNOWLEGE  ORGANIZING & REORGANIZING
	PROBLEM SOLVING / SOLUTION SEARCH
	NATURAL LANGUGE PROCESSING

    MODEL:
	TARGET OBJECT / OBJECTS & RELATIONS SET /
	OBJECT DESCRIPTION ON KNOWLEGE LEVEL	
	OBJECT PRESENTATION / IMAGINATION	
	OBJECT BEHAVIOR	/PREDICTIVE EQUATIONS ETC/
	OBJECT SYMULATION (SPEED)  & PROJECTION
	MODEL APPLICABILITY RECOGNITION 	


    MODELS OF MODELING :
	MODELS-OBJECTS-RELATIONS MODEL
	MODEL APPLICATION MODEL
	MODEL INSTANTIATION IN OBJECTS
	MODELING THE PERCEPTION


    GENERAL MODELS:
	OBJECTS - CHANGES - CASUALITY 
	OBSERVATION - ANALYSIS - ACTION
	SPACE AND TIME
	NEEDS - DESIRES - GOALS
	SELF- WHATBELONGSTOME- INTERFACE- WORLD

You can see that it is a limited task, although not a small one.

If such core SAI is developed, it can be incorporated into any of existing and/or new GAI projects, as well as augmented with any elaborate interfaces desired.


{{
SEED Technology

Integrated Brainstorming, Design, and Development 


}}

{{
Plan-Search-Remember-Learn-Generalize-Instantiate operation: second search gives lower priority to unsuccessful branching, 
gives higher priority to successful branching; uses additional search rules 
}}

{{
Search Related knowledge: decisions and their efficiency remembered

Logical Equations and solving them
Problem presentation as a Logical Equation
Search for a Solution

Splitting Problem and Integrating solutions
}}

{{
Natural language processing  => translation of notional referenses and sentnces, 
             into actionable internal references and relations, analogies etc.

multiaspect - multidimentional analogies
detailness of analogy
precision of analogy
level of analogy
{
unification of knowledge/data   => analogy linking
       at solution search time | at knowledge optimization time
approximate unification, quality and precision of analogy
}
{
   changes recording:
          what_context::by_which_processing::what_condition::which_variant
}
multistep rules for game against unintended consequenses.
do-nothing/skip step/
Scheduling/rescheduling  of moves

{{
referenses, dereferencing  :: level of
search
indexing, caching
common memory support
common knowledge support
[common] devivative knowledge search support
[common] decision support
infrastructure for agents and robots
}}

{{
specialized prosessing:
mass data processing,
mass data-stream processing,
imaging processing, 
communication processing,
math-expressions processing
math processing
}}

}}

GENERAL ARCHITECTURE

The things that are common in any intelligence:
integration of activity
self-application
self-development
self-control
Actual Thinking Operations


brain is a processor of different architecture, with different operation set
do we want to use brain architecture, if so at what level?
To have intelligence, we need to perform Thinking Operations, like analogy...
So they must be implemented using capabilities of specific processor, it is necessary and sufficient.

as we are going to create intelligence software, that is capable and doing self development, its self-development has to be software development.
we know how to do human software development, and we use IDEs for that
so AI-brain has to have IDE as a part of its structure.
The way that SWdeveloment is being done is OO, with the objects mapped into the memory and procedured.
the natural, "obvious" way is to keep things as they already have evolved where possible, as that is result of a long and successful process of understanding and optimisation of architecture and of development process.



{
     Memory objects vs Storage objects
     VirtualObject memory, alike Virtual memory with dirty flags etc

     Memory pointer links vs Symbolic name links
}