Meta-Knowledge Engineering & Management Server ,                     ENEA (Ente Italiana per la Ricerca, Energia e l'Ambiente)
Short  e-paper .  References: Adam Maria Gadomski, ENEA's e-paper, http://erg4146.casaccia.enea.it/wwwerg26701/gad-dict.htm, page since 1999,  at present on the MKEM Server.

  Meta-Ontological Assumptions: 
 Information, Preferences and Knowledge

  universal interrelations (cognitive IPK  architecture) 


 The TOGA proposal of  universal paradigms for Knowledge Engineering & Management and, for necessary for it, intelligent agent modeling, begins at  the cognitive IPK model considered as Universal  Reasoning Architecture Paradigm (URAP). It is funded on the (goal-oriented) selection, re-definition and systemic socio-cognitive application of a few canonical terms which enable to define the generic kernel of every symbolic reasoning function. The main original idea relies on  the generalization of well known data processing and calculation algorithms conceptual frame where, for an assumed application domain D, an algorithm or another processing expression is considered a knowledge, data are information and individual preference relations are used for the choice of a proper knowledge expression. It means, in the case of models/algorithms in physics we use the following calculation frame:

 

                           Information' D= KnowledgeD( Information D), 

 

which is valid in a domain D where the choice of the KnowledgeD depends on the objective/goal/purpose of the intervention of a physicist/mathematician/engineer, it means on his/her preferences related to the state of D.

 

The I,P,K (Information, Preferences, Knowledge) conceptualization was mentioned  in economy problems in an intuitive and very synthetic manner by F.A. Hayek:

    “If we possess all the relevant information, if we can start out from a given system of preferences, and if we command

     complete knowledge of available means, the problem which remains is purely one of logic.”    ( Hayek,1945 *)

– It means, the “logic” is the way how to use them together, and, we may add, the efficacy of IPK processing is exactly the fundamental “task” of intelligence (Gadomski, 1993).

 On the other hand, the modeling of cognitive decision-making has been latter obscured rather than illuminated by many of subsequent refinements of cognitive theory,  where usually some arbitrarily chosen cognitive properties are abstracted from complex contexts without  sufficient argumentation and aggregated to the form of numerous partial and reciprocally independent models.

-----------------

(*) F.A. Hayek, The Use of Knowledge in Society , The American Economic Review, Vol. 35, No. 4 (Sep., 1945), pp. 519-530. http://www.virtualschool.edu/mon/Economics/HayekUseOfKnowledge.html

 


 Definitions of the basic concepts of the IPK
interrelations according to the TOGA meta-theory standard ( see also  socio-cognitive engineering paradigms), are the essential part of  the assumptions of the TOGA  meta-ontology and epistemology (Gadomski, 1993) and can be considered

 

            Data: everything what is/can be processed/transformed in computational and  mental processes. 
                     Concept data is included in the ontology of "elaborators", such as developers of methods, 
                     programmers and other computation service people.
                     In this sense, data  is a relative term and exists only in the couple (data, processing) or

                     (data, calculations).

 

            Information, I: data which represent a specific property of the domain of human or artificial agent's
                     activity (such as: addresses , tel. numbers,  encyclopedic data, various lists of names and
                     results of  measurements). 
                     Every information has always a source domain. It is a relative concept.
                     Information is a concept from the ontology of modeler/(problem solver)/decision-maker.

 

            Knowledge, K: every abstract property of human/artificial agent which has ability to process/transform
                      a (quantitatively / qualitatively) information  into other information or in another knowledge
.     

                      It can be: instructions, emergency procedures,  exploitation/user manuals, scientific materials,

                      models and theories, as well as learning and teaching   methods and methodologies.
                      Every knowledge has its reference domain where it is applicable. It has to include the source 
                      domain of the processed information. It is a relative concept. For more.
 

            Preference, P: an ordered relation among two properties of the domain of activity of a cognitive agent,
                     it indicates a property with higher utility. Preference relations serve to establish an intervention 
                     goal of an agent.
Cognitive preferences are relative .
                     An agent ( preferences agent) which manages  preferences of an intelligent agent  A, can be an external  
                    agent or  the internal part of the agent A.. Reasoning/inference  on preferences are critical for criteria
                    building/choice in a cognitive decision-making.
 
            Goal, G: a hypothetical state of the domain of activity which has maximal utility in a current situation.
                     Goal serves to the choice and activates proper  knowledge which process new information.

            Document: a passive carrier of  Knowledge,  Information and/or Preferences (with different 
                    structures), comprehensive for humans, and it has to be recognized as valid and useful by 
                    one or more human organizations, it can be physical or electronic. 

            Computer Program:
                   - from  the modelers and decision-makers perspective: an active carrier of different structures
                   of  knowledge expressed in computer languages and usually focused on  the realization of predefined
                   objectives (its design-goal). It may include build-in preferences and information and/or request
                   specific IPK as data.
                  - from the software engineers perspective: a data-processing tool ( more precise technical def. you 
                    may find on the Web).

             The above presented concepts consist of a computational entity discussed more in deep in
             the white paper: Personoids Paradigms; http://erg4146.casaccia.enea.it/wwwerg26701/per-para.html 

 

             Utility and efficacy of the IPK bases depend on their many meta-properties, such as: wisdom   

             and belief.
 

            Wisdom (TOGA definition),  possible property of a preferences system, it is a  meta-attribute of the set
                     of   preferences enabling high efficacy in the achiving of preselected goals  in the preselected class
                     of  domains-of-activity of  an intelligent agent.
                     Wisdom (as an indicator) has a qualitative value domain, its study are included in such domains as
                     Axiology, Ethics and
Business. In the case of artificial intelligent tools,
 it is investigated by quality and

                     knowledge engineers.

            

             Beliefs  represent user's relation to particular data. They can relate to  information, knowledge
                     or  preferences  when they are used  without sufficient  verification or validation.
                      Belief is a meta-attribute, therefore it represents  relative point of view on the contens of IPK bases ,
                     and   depends on the particular/current observer's IPK (an intelligent agent too) .
                      For example, in a concrete situation, what is a belief for an intelligent agent A can be validated
                      knowledge  for an intelligent agent B.
            In the real world situations, in decision-making, every  I, P and K entity has the certainty attribute,
            and if its value is not sufficiently high but the entity is used by decision-maker then it can be also
            called his/her/its belief.

          


  First Remarks

 

   Remarks I:  Information and knowledge are relative concepts and what is called knowledge can be  considered
            as information on a higher, more abstract meta-activity level (meta-reasoning level). 

 

  Remarks II: The IPK ontology requires a generic assumption on the existence of an abstract intelligent agent
            (AIA1 , Abstract Intelligent Agent, 2,  Abstract Intelligent Agents newsgroup) and its/his/her domain of activity
          . (see also here)

 

  Remarks III: From the TOGA's intelligent agent centered perspective, information, preferences and  knowledge
             are only comprehensive if they are related to the agent's domain-of-activity before pre-selected.
          

  The IPK ontology framework consists of the paradigm of personoids architecture

 

  Context remarks, another similar approaches
        - In less formal than TOGA theories (recently under continuous development), such as: "Information, Knowledge, 
           Wisdom" and "Data, Knowledge, Wisdom" (see Web) the concepts of information and knowledge are not 
          sufficiently computationally defined, and wisdom is also used in a metaphoric sense.
           In TOGA, wisdom is a specific property of  the preferences rules system of an intelligent agent ( see above).

         - We may expect that  intuitive conceptualization frameworks as: (Intelligence, Knowledge, Information), 
        (Knowledge, Intelligence, Wisdom) or ( DIKW ) will be slowly transformed (2004) to the form of  the TOGA's IPK 
       architecture/ontology (maybe with other names but  with the  same definitions - unfortunately, people do it).
         By the way, the same  evolution process is seen  in the   numerous variants of the BDI  (Believe, Desire, Intention)
       models of  various cognitive agents (see Web). 
        Another similar approach, developed independently, in parallel to TOGA, is SOAR (proposed by Allan Newell as a base
        for his General Intelligence). It is now an advanced AI development environment  but the  concepts of  information
        and knowledge are not clearly distinguished  (2004).
        - It is yet useful to notice that for software system developers which start from before developed models, quasi all 
        problem specification information are considered as data, therefore the IPK conceptualization  is not well visible from 
        such technical perspective.


  More about meta-ontological and epistemological TOGA assumptions:
MKEM Home
Ontology
Knowledge structure
IPK architecture
TOGA Meta-theory

  Google search 
        ( Nov. 16, 2004):   Information, Preferences, Knowledge    2.250.000 docs. 1st & 2nd  from MKEM server
        ( Sept. 30, 2005):  Information, Preferences , Knowledge   17.100.000 docs. 1st & 2nd  from MKEM  server
        ( Sept. 30, 2005):  Information, Preferences             103.000.000 docs. 1st is this page.

Go to A.M.Gadomski Home-page                       All comments are welcome.


© 1997,2007, ENEA. HID, A.M.Gadomski. All  stuff  may be freely distributed in its full original format, if selected ideas are to be published in another format, a reference to the original source and to the author must be evidenced, intellectual rights reserved. No permission is granted to download and save professional images or any other material from these pages other than for viewing and citation purposes.  These are ENEA's research pages,  they represent the opinions of the contributors, but not necessarily of ENEA. (last small comments added - 23 May 2008)