Applying Multi­Level Typing to Model Knowledge­Intensive Processes Previsão de defesa em Abril/2019

Abstract

  Modeling   Knowledge­Intensive  Processes  (KIP)  is  very   important for   understanding   critical   scenarios   in   current   organizations.   KIPO (Knowledge­Intensive   Process   Ontology)   is   an   ontology   well­founded,  semantically   rich   conceptualization   of   KIP.   However,   it   is   difficult   to distinguish   instances   and   models   in   KIP.   Our   goal   is   to   propose   an application   with   the   notion   of   multi­level   conceptual   modeling   for representing elements with multiple classification level.

1. Introduction

However, [Franca et al. 2014] observed the difficulty of distinguishing instances and   models   in   KIP.   There   are   many   problems   in   modeling   Knowledge­Intensive Process,   mainly   because   it   is   very   common   to   misunderstand   instances   and   models when thinking about the characterization of each element as part of a KIP definition.

When   this   situation   is   solved,   the   alignment   of   the   Knowledge   Management strategy to the Business Process Management strategy of an organization will facilitate the identification of faults, the correction of errors and adaptation to changes. The more information the managers of organizations obtain, the better decisions will be made.

For   address   this   problem,   we   propose   to   apply   the   notion   of   multi­level conceptual   modeling   [Carvalho   et   al.   2017]   for   representing   elements   with   multiple classification  levels, such as MLT (Multi­Level Theory), which is a theory  of multi­ level modeling for capturing and analysing a number of nuances related  to modeling [Carvalho and Almeida 2016]. Moreover,   we will also use the powertype pattern for representing KIP characterizations. 2.

Related Work

There   are   many   kinds   of   multi­level   modeling   which   the   main   goal   is   to   identify concrete and abstract elements . One of those elements is the Materialization [Goldstein and   Storey   1994],   which   means   the   relationship   between   two   entity   types,   one   that represents a conceptual object, for example, a TV Model, and the other that represents its corresponding concrete object, in this case, actual Tv Sets. This strategy  helped us to determine   which   elements   are   concrete   existents   in   KIPO.   Another   multi­level modeling   approach  is  Powertype  [Odell  1994], which  defines  its  concept   informally using   regular   associations   between   the   powertype   and   a   base   type.   Based   on   this concept, it is possible to build constructs  to denote cross­level relations between types defined   in   MLT.   This   concept   is   used   in   our   research   for   supporting   i   such  characterizations in models.

 There   are   other   kinds   of   multi­level   modeling   with   the   focus   on   reducing  accidental   complexity   in   models,   for   instance:   Deep   Instantiation   [Atkinson   and Künner   2008]   ,   Dual   Deep   Instantiation   [Neumayr   and   Schrefl   2014],   Melanee [Atkinson   2012],   among   others.   All   of   them   are   also   very   useful   to   address   the shortcomings   of   the   UML­based   model.   However,   they   are   not   proper   for   the instantiation of Knowledge­Intensive Process.  These techniques are not usually applied in unstructured processes. 

[Carvalho et al. 2017] presented a theory for multi­level modeling called MLT, which distinguished between types (that have other types as instances)  and individuals (cannot be instantiated anymore). The notion of type order is used in MLT. Carvalho also   combined   UFO   (Unified   Foundational   Ontology)   [Guizzardi   2005]   with   MLT (Multi­Level Theory) [Carvalho et al. 2017] for establishing a hierarchy of conceptual models, where the concepts of UFO instantiate and specialize elements of MLT, thereby respecting MLT’s axiom and leveraging the use of structural relations and MLT pattern in UFO. Many rules applied in this combination were also adapted to be used in our research, because KIPO is well­founded in UFO. 

Metodology

To achieve the goal of this research, it is necessary the following phases: 2. Combine MLT theory with KIPO:  In this phase, all elements KIPO will be reviewed under   MLT Theory. Based on the combination created in [Carvalho and   Almeida   2016],   the   concepts   of   KIPO   will   instantiate   and   specialize elements of MLT.   These concepts in KIPO’s taxonomy of individuals will be instances of “1stOT” specializing “Individual”. The concepts in the taxonomy of “type” are instances of   “2ndOT” specializing   “1stOT”.   We will use UML to provide   a   graphical   representation,   because   it   is   a   well­known   pattern   by modelers. To exemplify this combination, Figure 1 shows one of the main KIPO elements   (KIPCO::Agent  with   their  specializations  KIPCO::Innovation   Agent and  KIPCO::Impact Agent)  combined  with MLT (Theory Multi­Level),  using Powertype   concepts   [Odell   ].   Each   element   has   a   meaning,   for   instance: KIPCO::Agent is a participant of process which has his/her actions motivated by his/her   intentions,  KIPCO::Innovation   Agent  is   responsible   for   incorporating innovations   in   an   knowledge­intensive   activity   and  KIPCO::Impact   Agent  is responsible for executing a KIP and identifying questions during   execution of KIP.   KIPCO::Agent Type  is instance of  “2ndOT” (second order),  specializes “1stOT”   (first   order)   and  characterizes  KIPCO::Agent,   which  is   instance   of “1stOT” and specializes “Individual”. According [Odell  ] the specializations of KIPCO::Agent   (KIPCO::Innovation   Agent   and   KIPCO::Impact   Agent)  are instances of KIPCO::Agent Type. 

The Figure 2 exemplifies the application of MLT with KIPO (cited above) using a scenario of the participants of a dissertation assessment. We used stereotype [Sellers and   Perez   2005]   for   identifying   each   element   in   diagram   that   is   referencing   the application MLT to KIPO.  3.

Adapt modeling tool with combination: It is necessary to automate the process of instantiation. In this phase it will be implemented the combination in a modeling tool, which will have to read a knowledge base (owl, xml, etc.) and will represent, through of UML diagrams, instances and models found in these bases. 4. Execute Case Study and Experiment:  Apply the tool (already adapted) in knowledge base with different contexts (tickets of call center, governamental wiki and others) . The diagrams will be evaluated by specialists in data modeling experts in multi-level theory, besides users of the knowledge bases.

Expected Outcomes and Results

The   main   outcome   is   KIPO   reviewed,   according   to   the   MLT   theory.   Another contribution   is   the   deepening   of   the   discussion   about   instances   and   models   in   the domain of KIP.

The   expected   results   with   the   new   version   of   KIPO   is   the   possibility   of distinguishing instances and models. A practical example is, considering the domain of elaboration   dissertation,   universities   need   to   manager   the   types   of   “Knowledge Intensive Activity” (KIA) (“Select Advisor”, “Define Problem”) that are executed. They may need to classify those KIA types giving rises to types of KIA types. In this case, “Select   Advisor”   could   be   considered   as   examples   “Dissertation   Enrollment   Process Type” and  “Define Problem” which is an example of “Define Problem Type” . Finally, they need to track the specific activity of  problem definition of a student’s dissertation (e.g.   “Define   Problem   of   Mary”).   So,   for   representing   this   case,   we   need   to   use differents classification  levels,  such as individual KIA (“Define Problem of Mary”) , KIA type (“Define Problem”) , and types of KIA Type (“Define Problem Type”).

Final Considerations

In   this   paper,   we   argued   about   the   importance   of   Knowledge­Intensive   Process   and how it is a valuable resource into organizations. We also discussed that it is difficult to manage   a   KIP   and   how   the   ontology   KIPO   helps   to   understand   it.   However,   the problem identified r was the difficulty in distinguishing instances and models in KIP, which KIPO does not  address in its current form. Therefore,   a solution was proposed based on the application of combination MLT and KIPO, using concepts of powertype, defined by Odell. 

The results expected are  to represent, through of UML diagrams, instances and models,   from   reading   a   knowledge   base,   using   the   strategy   defined.   These representations will be generated by an implemented tool.  [Odell 1994] Odell, J.(1994): Power types. In: Journal of Object­Oriented Programing, 7(2), pp. 8­12. [Guizzard   2005]   Guizzard,   G.(2005):   Ontological   Foundations   for   Structural Conceptual Model. In: Enschede, The Netherlands. Telematica Institut Fundamental Research Series, No. 015 (TI/FRS/015). Netherlands [Sellers and Perez 2005] Sellers, B.H. and Perez, C.G.(2005): Connecting Powertypes and   Stereotypes.   In:   J   OURNAL   OF   O   BJECT   T   ECHNOLOGY.Online   at http://www.jot.fm. Published by ETH Zurich, Chair of Software Engineering ©JOT, 2005. Vol. 4, No. 7, September ­ October 2005. Zurich [Maldonado   2008]   Maldonado,   M.   (2008):   Impact   analysis   of   knowledge   intensive process  creation   and  transfer   policy:  a  system  dynamic   model.   M.Sc.   dissertation. Programa   de   Pós­Graduação   em   Engenharia   e   Gestão   do   Conhecimento,   UFSC, Brazil (in Portuguese) [Neumayr and Schrefl 2014] Neumayr, B. and Schrefl, M. (2014): Abstract vs Concrete

Clabjects in Dual Deep Instantiation. In: Proc MULT 14 Workshop, pages 3­12. [Carvalho and Almeida 2016] Carvalho, V. A. and Almeida, J. P. A.(2016): Toward a well­founded   theory   for   multi­level   conceptual   modeling.   .   Software   &   Systems Modeling, Springer Berlin Heidelberg 

References

[Goldstein and Storey 1994 ] Goldstein, R.C. and Storey , V. C. ( 1994 ) : Materialization. In: IEEE Transactions on Knowledge and Data Engineering  (Volume: 6 ,  Issue: 5, Oct 1994 ).

[Atkinson and Künner 2008 ] Atkinson, C. and Künner , T. ( 2008 ): Reducing accidental complexity in domain models . In: Software & Systems Modeling .July 2008 , Volume 7 , Issue 3, pp 345 - 359 .

[Atkinson 2012] Atkinson, C. ( 2012 ) : Melanie - Multi-level Modeling and Ontology Engineering Environment . In: Proceeding  MW '12  Proceedings of the 2nd International Master Class on Model-Driven Engineering: Modeling Wizards. Article No. 7 . Austria

[França   et   al.   2014 ]   França,   J.B.S. ,   Netto,   J.M.S. ,   Carvalho,   J.E.S. ,   Santoro,   F.M. , Baião ,   F.A. ,   Pimentel,   M.( 2014 ) :   “KIPO:   the   knowledge­intensive   process ontology” . In: Software &  System  Modeling. Springer.