Mapping ontologies and contexts: from theory to a case study. Dmitry Kudryavtsev1 Abstract: Ontologies are used actively as a knowledge communication. In order to solve these problems it is necessary to representation, retrieval and navigation tool to improve knowledge define context and make explicit mapping between content sharing, exchange and communication. In order to provide ontology (or knowledge resource directly) and context. In the paper effective communication ontologies should be mapped with the [Section 2] describes existent approaches to a context definition context. This paper analyses existing approaches towards the very and mapping context and ontologies. [Section 3] marks out two definition of context and suggests two context types. Requirements context types and suggests the requirements for effective for effective knowledge representation based on two context types knowledge representation with respect to these types. [Section 4] and on mapping ontologies and context are suggested. These initiates case-study and describes real-life knowledge requirements are considered and factored in the following case communication task and corresponding problem. Knowledge study by consecutive mapping different context types and content Navigator (KN) is suggested as the solution for this knowledge ontology . This case study describes Knowledge Navigator – a map communication task. [Section 5] suggests KN framework and brief that relates contents of Formalized Management methodology with description, which satisfies requirements from [Section 3] and is the corresponding context in order to reach effective knowledge based on a consecutive mapping different context types and content communication to end users. ontology. 1 INTRODUCTION. 2 DEFINITION OF CONTEXT AND Nowadays organizations implement special tools and technologies RELATED WORK to share, exchange and communicate knowledge. In order to be In [1] it is suggested to focus on the context as highly relevant for effective, these tools and technologies must provide users with retrieval within an organization. In modeling the context the relevant information in due time without being flooded with authors deal with two issues: irrelevant data. To support the sharing and exchange of knowledge • the intended application context of a knowledge item, and both among information systems and people it is useful to define • the context a knowledge item was created in. ontology [6]. Now ontologies are already employed in portals, The Authors suggest that information context be expressed in terms corporate memories, e-commerce and other knowledge of the organizational structure and the process models. These in management systems (see [1], [2], [11]). With respect to human- turn are expressed in terms of the enterprise ontology. The design computer interactions ontology often works as a representation, of the enterprise ontology is built on insights and developments retrieval and navigation tool. In playing such a role ontology from the enterprise modeling, business process modeling, and usually specifies the Content of knowledge resources. Such an organizational modeling in knowledge-based systems [13]. In [2] a ontology can be called Content ontology. similar approach is used for semantic mapping between the sellers’ There are two problems that render the usage of Content ontology supply and buyers’ needs at an electronic knowledge market. less efficient. Similar enterprise ontology oriented approach to the context 1. A Content ontology user is unable to set links between his/her definition can be also found in the knowledge mapping task, problem, situation and notions in the Content ontology, thus technologies [4], [14]. he/she is unable to transform information into action. The definition of context described above resulted from the “In many situations a content ontology user may not know the knowledge management field, whereas in the semantic web field details of a solution, but he knows the details of his problem” [2]. there is another useful definition of the context. “One of the fundamental tenets of knowledge management is that According to [3] Ontologies are shared models of some domain knowledge must link to and improve business processes. Without a that encode a view which is common to a set of different parties map of the processes, goals, and knowledge assets inside one’s Contexts are local (where local is intended here to imply not organization, it will be difficult to reach one’s destination.” [14] shared) models that encode a party’s view of a domain [7]. 2. A Content ontology user is unable to match his/her personal The authors argue that an ontology is contextualized, or that it is a mental model with notions in the Content ontology because of contextual ontology, if it is kept local (and therefore not shared semantic and syntactical specialties of a person and ontology- with other ontologies) but its contents are put in relation with the creator. contents of other ontologies via explicit mappings. This mapping This problem is taken from an elaborated field of semantic web provides syntactic and semantic interoperability and deploys a where it is known as a mismatch between ontologies (see [5], [9]) variety of methods, coming from very different areas. They (it is suggested to use analogy between ontology and personal include: linguistic, statistical, structural and logical methods (see mental model in the paper). [5], [8], [9]). All these problems are related with the notion “context”. These problems make problematic effective knowledge sharing and Saint-Petersburg State Polytechnic University, Russia BIG-Petersburg, Russia, dk@big.spb.ru 3 MODEL OF CONTEXTS AND contents) and divided into topics (content ontology nodes) it is REQUIREMENTS FOR EFFECTIVE rather hard to communicate it because the way the methodology can be used, its potential users and the methodology itself have KNOWLEDGE REPRESENTATION their own specialties. These specialties can be considered as Resuming Section 2 there are two main definitions of context that communication problem and are as follows: affect communication problems (Section 1): a. Different organizations that intend to use the methodology face Def 1. Context model reflects: different problems and tasks. Many problems and tasks do not • the intended application context of a knowledge item, and require usage of every topic of methodology. • the context a knowledge item was created in. b. Implementing such a methodology is not a task faced by one and is expressed in terms of the enterprise ontology. person or a small group only; it requires a joint effort made by Def 2. Contexts are local (not shared) models that encode a party’s many persons employed in the organization. As a result the view of a domain. target audience for the methodology implemented is very broad In order to distinguish types of context and set requirements for and involves many people in a management activities oriented knowledge representation working definitions are suggested for organization (ranging from directors’ boards to linear every type of context. The first working definition is based on a managers). It is a subset of topics that is to be read and learned semiotic model [10]. Traditionally the semiotic model includes: by a majority of users’ categories. c. The core of the methodology integrates words quite unusual • Syntax which reflects rules and relations between signs of and new for the majority of Russian managers (Corporate / any language Enterprise Architecture, Business Engineering). In addition • Semantics which reflect relations between signs and their management research and practice have no conventional terms meaning and concepts. Thus words and phrases used in the methodology • Pragmatics which reflect relations between signs and their and especially in the topic headings can be misunderstood and users and creators users will be unable to set a relation between their mental This model together with Def 1 makes possible to consider the models and topics of the methodology. context in Def 1 as pragmatic context. In order to effectively communicate methodology with respect to The Context in Def 2 will cover all the components of the semiotic the specialties described above Knowledge Navigator (KN) was model making it impossible to define it uniformly in terms of a created. semiotic model. Thus the context in Def 2 will be termed and used in this paper as local context. Pragmatic context can be either shared or not. Consequently the 5 CASE STUDY: KNOWLEDGE NAVIGATOR former is represented by ontology and the latter is by a set of local contexts. FRAMEWORK AND DESCRIPTION The requirements for effective knowledge representation which Input data for KN are content ontology and the very content. provides for a solution of the problems from Section 1 are as In order to satisfy the requirements for effective knowledge follows: representation KN – end-user solution – integrates three tools Requirement 1: Every ontology must be either shared by all the (Figure 1): communication participants or be mapped with corresponding local contexts of every participant (group of similar participants). Local Task&Problem Context n Task Requirement 2: Every knowledge resource must be mapped with a Task- Local Task&Problem Context 2 Context oriented Local Task&Problem Context 1 ontology Formalized pragmatic context (either directly or by means of the content management navigator ontology). methodology These requirements are further considered and factored in the case Activity Role study. Role- Context Context Content Content ontology oriented ontology ontology navigator 4 CASE STUDY: TASK SETTING AND PROBLEM DESCRIPTION Local Content Context n Local Content Context 2 Formalized management methodology (“methodology” further in Semantic Local Content Context 1 the paper) is a product of the management consulting company navigator BIG-Petersburg. This methodology is initially presented in the form of a book, but the concept “Formalized Management Figure 1. Knowledge Navigator Framework Methodology” is used due to the plans of application of other media, such as e-books or knowledge portals. 1. Task-oriented navigator (“What for” – navigator) This methodology reflects the experience of consultants gained It helps users to choose topics to solve certain tasks and during business process improvement and restructuring of problems of organization. organizations in Russia and CIS countries. This navigator maps content ontology with Pragmatic context, The goal of this methodology is to help different organizations in which is represented in the form of Task Context ontology. But solving their managerial problems and improving levels of although the latter ontology results from the analysis made by a management. Thus the main objective is to provide each potential consulting company and is shared by the authors, it is not organization based user of the methodology with necessary shared by prospective users and consequently does not satisfy knowledge to help realize the tasks and functions they face. Requirement 1 from [Section 3]. In order to help the users In order to achieve this objective the methodology must be identify their local problems every node in Task Context effectively communicated to its potential users. Although ontology is mapped with a set of descriptive local task and methodology is well-structured with a content ontology (=table of problem contexts of users. These local contexts are given even in user linguistics. Finally users of this navigator do two Table 2. Role-oriented navigator – example consecutive mappings, see Step 1 and Step 2 in Figure 2. Activity Context Role Context Impor- Content Ontology ontology ontology tance /Topics Perform external Ideology of modern Task Content ~ Local Task&Problem Context ontology and internal analysis organization Director of Context ontology Develop business Business Business Engineering Step 1 { strategy Development and modeling Develop and set Corporate Architecture U organizational goals as a control object Step 2 3. Semantic navigator (“What about” – navigator) This navigator helps users to relate topics in authors language with their knowledge and thus refine a subset of topics to learn. This navigator maps the Content ontology with the Local Figure 2. Task-oriented navigator - two consecutive mappings Content Contexts, which are represented by the keywords. Real-life example for shaded blocks from Figure 2 is Namely this combination of 3 tools together with internal mapping represented in Table 1. will provide effective communication. Such a framework of KN takes into account knowledge communication specialties (problem) Table 1. Task-oriented navigator – example from [Section 4] and satisfies the requirements from [Section 3]. Local Task&Problem Context Task Content Impor- Context Ontology tance ontology /Topics 6 CONCLUSIONS 1. You might have encountered Business situations of complete chaos { Engineering This paper suggested the requirements for effective knowledge resulted from disorganization in and modeling representation based on mapping ontologies and context with your company. These cause the Corporate respect to two types of the latter. It described a solution for real-life same problems to reoccur. To Architecture knowledge communication task called Knowledge Navigator. This ---------------------------------------- establish U as a control solution illustrated consecutive mapping ontologies and contexts – 2. The strategy issues are left order object unheeded in your company. The mapping which was necessary to effectively communicate Tools of knowledge to different users, which solve different tasks and have main question your company managers are faced with is “how ~ Business different understanding of domain and background. to cater to the clients’ order” Engineering Importance: ~ Critical { Important U Useful 7 REFERENCES 2. Role-oriented navigator (“Who” – navigator) [1] A. Abecker, A. Bernardi, K. Hinkelmann, O. Kuhn, M. Sintek, Toward It helps users to choose topics for learning with respect to their a Technology for Organizational Memories IEEE Intelligent Systems. – Roles in the organization. 1998. - №3, 40-48. This navigator maps content ontology with Pragmatic context, [2] A. Abecker, D. Apostolou, W. Maas, G. Mentzas, C. Reuschling, S. which is represented in the form of Role Context ontology. Tabor, Towards an Information Ontology for Knowledge Asset Trading Similarly to task-oriented navigator, Role Context ontology is Presented at the ICE 2003 - 9th International Conference of Concurrent Enterprising, Espoo, Finland, 16-18 June 2003 ambiguous and polysemantic for the users, because Roles [3] P. Bouquet, F. Giunchiglia, F. Harmelen, L. Serafini, H. (nodes of Role Context ontology) can bear different Stuckenschmidt, Contextualizing Ontologies, Journal of Web Semantics, responsibilities in different organizations. Thus the Role 2004, Vol.1, №4. Context ontology is mapped with the elements derived from [4] M. Eppler, Making Knowledge Visible Through Intranet Knowledge the next Pragmatic context - Activity Context ontology. The Maps: Concepts, Elements, Cases Proceedings of the 34th Hawaii Activity Context ontology can be considered as shared by International Conference on System Sciences - 2001 potential users, because all the management activities presented [5] F. Harmelen, Ontology Mapping: A Way Out of the Medical Tower of are typical for different organizations. Finally users of this Babel? AIME 2005, pp. 1–4, 2005. [6] T. Gruber, A translation approach to portable ontology specifications. navigator also do two consecutive mappings, see Step 1 and Knowledge Acquisition, 1993, Vol. 5, 199- 220. Step 2 in Figure 3. [7] C. Ghidini, F. Giunchiglia, Local models semantics, or contextual reasoning = locality + compatibility, Artif. Intell. 127 2 (2001) 221–259. Activity Role Context Content [8] F. Giunchiglia, P. Shvaiko, Semantic Matching. In The Knowledge Context ontology ontology Engineering Review Journal, vol. 18(3), pp. 265-280, 2003. ontology [9] M. Klein, Combining and relating ontologies: an analysis of problems and solutions, Workshop on Ontologies and Information Sharing, IJCAI'01, 2001, №4-5. Step 1 [10] C. Morris "Foundations of the Theory of Signs." International Step 2 Encyclopedia of Unified Science, ed. Otto Neurath, vol. 1 no. 2. (Chicago: University of Chicago Press, 1938. Rpt, Chicago: University of Chicago Press, 1970-71). [11] S. Staab, A. Maedche Knowledge Portals: Ontologies at Work. AI Figure 3. Role-oriented navigator - two consecutive mappings Magazine 2001, Vol. 22, №2, p. 63-75. [13] M. Uschold, M. King, S. Moralee and Y. Zorgios The Enterprise Ontology AIAI, The University of Edinburgh, 1997. Real-life example for shaded blocks from Figure 2 is [14] W. Vestal Knowledge Mapping: The Essentials for Success APQC: represented in Table 2. Publications. 2005.