Managing Semantic Web evolution in the Organization...

6
Managing Semantic Web evolution in the Organization (SWO) Mohamed BINIZ, Mohamed FAKIR Laboratory of Information Processing and Decision Support (TIAD), Faculty of Sciences and Technics,University Sultan Moulay Slimane, Beni Mellal,Morocco. Email:[email protected], info[email protected] Abstract —Recently, similarities between database-schema evolution and ontology evolution allow us to build extensive schema evolution. However, there are also important differences between database schemas and ontologies. The differences restrain in different usage paradigms, the presence of explicit semantics and different knowledge models. These differences have important implications for the development of ontology-evolution frameworks: The traditional distinction between versioning and evolution is not applicable to ontologies. The set of change operations for ontologies is different. We must develop automatic techniques for finding similarities and differences between versions Although, there are no sophisticated methods available yet to support all the aspects of change management for ontologies, it surely is an active research field. Most of the work has been done under the titles of Ontology Evolution and Versioning. Keywords: Ontology, semantic web, XML, RDF, RDFS, OWL, OWL 2, Evolution, Schema. I. INTRODUCTION The invention of the Internet is to share all the data generated with everyone in the world, which has led to a huge change (data, technology...). Despite many advantages, this has also given rise to the problem of extracting relevant information about the huge amount of data that we face on a daily basis. But the problem lies in how to share such knowledge, acquired, these services and applications such as those experiences? These knowledge assets become the competitiveness and survival of many organizations because they must respond to frequent changes in their field activities. In this context of rapid change and development, the establishment of an institutional memory is required for the organization to achieve these Aims. In the next generation semantic web, the idea is to model the content of Web resources by adding semantics as metadata to make it understandable by machines resources according to the W3C founder Tim Berners-Lee. [1] This idea is better cooperation between humans and machines, collective memory can be modelled as a semantic web organization (SWO), which includes heterogeneous changing distributed on all network resources and indexed using annotations semantics expressed with the vocabulary provided by shared ontologies. However, organizations are in dynamic and changing environments due to changes in their objectives, technologies and processes ... These changes in the real world often lead to the need for changes SWO. Example: changes in the environment of the organization lead to continuous changes in the conceptualization of a field of activity that could affect the underlying ontologies. These changes in the underlying ontology must then spread to all the semantic annotations that are created on the basis of the common vocabulary of ontologies. Consequently, these changes affect the operations and performance of SWO system. In this work, we focus on the ontology evolution domain, we propose an ontology that describes this field and contributes to the development process itself, because in our view, the evolution of ontology is a following instances of ontology evolution of a SWO. This sequence of instances will be stored in a change log. This paper is organized as follows: section 1 describe the semantic web and related technologies, section 2 presents the concept of evolution and ontology evolution approaches existing in the literature. Section 3 is devoted to the tools supporting the development of ontology, in Section 4 we will discuss about SWO approach. Finally, we conclude.

Transcript of Managing Semantic Web evolution in the Organization...

Page 1: Managing Semantic Web evolution in the Organization (SWO)tal.ircam.ma/conference/data/papers/7.pdf · view, the evolution of ontology is a following instances of ontology evolution

Managing Semantic Web evolution in the

Organization (SWO)

Mohamed BINIZ, Mohamed FAKIR Laboratory of Information Processing and Decision Support (TIAD),

Faculty of Sciences and Technics,University Sultan Moulay Slimane, Beni Mellal,Morocco. Email:[email protected], [email protected]

Abstract —Recently, similarities between database-schema evolution and ontology evolution allow us to build extensive schema evolution. However, there are also important differences between database schemas and ontologies. The differences restrain in different usage paradigms, the presence of explicit semantics and different knowledge models. These differences have important implications for the development of ontology-evolution frameworks: The traditional distinction between versioning and evolution is not applicable to ontologies. The set of change operations for ontologies is different. We must develop automatic techniques for finding similarities and differences between versions

Although, there are no sophisticated methods available yet to

support all the aspects of change management for ontologies, it surely is an active research field. Most of the work has been done under the titles of Ontology Evolution and Versioning.

Keywords: Ontology, semantic web, XML, RDF, RDFS, OWL, OWL 2, Evolution, Schema.

I. INTRODUCTION

The invention of the Internet is to share all the data generated with everyone in the world, which has led to a huge change (data, technology...). Despite many advantages, this has also given rise to the problem of extracting relevant information about the huge amount of data that we face on a daily basis.

But the problem lies in how to share such knowledge, acquired, these services and applications such as those experiences? These knowledge assets become the competitiveness and survival of many organizations because they must respond to frequent changes in their field activities. In this context of rapid change and development, the establishment of an institutional memory is required for the organization to achieve these Aims.

In the next generation semantic web, the idea is to model

the content of Web resources by adding semantics as metadata to make it understandable by machines resources according to the W3C founder Tim Berners-Lee. [1]

This idea is better cooperation between humans and machines, collective memory can be modelled as a semantic web organization (SWO), which includes heterogeneous changing distributed on all network resources and indexed using annotations semantics expressed with the vocabulary provided by shared ontologies. However, organizations are in dynamic and changing environments due to changes in their objectives, technologies and processes ... These changes in the real world often lead to the need for changes SWO. Example: changes in the environment of the organization lead to continuous changes in the conceptualization of a field of activity that could affect the underlying ontologies. These changes in the underlying ontology must then spread to all the semantic annotations that are created on the basis of the common vocabulary of ontologies. Consequently, these changes affect the operations and performance of SWO system.

In this work, we focus on the ontology evolution domain, we propose an ontology that describes this field and contributes to the development process itself, because in our view, the evolution of ontology is a following instances of ontology evolution of a SWO. This sequence of instances will be stored in a change log.

This paper is organized as follows: section 1 describe the semantic web and related technologies, section 2 presents the concept of evolution and ontology evolution approaches existing in the literature. Section 3 is devoted to the tools supporting the development of ontology, in Section 4 we will discuss about SWO approach. Finally, we conclude.  

Page 2: Managing Semantic Web evolution in the Organization (SWO)tal.ircam.ma/conference/data/papers/7.pdf · view, the evolution of ontology is a following instances of ontology evolution

II. BACKGROUND WORK

1. SEMANTIC WEB AND RELATED TECHNOLOGIES

The current web is essentially syntactic structure of

resources being well defined, but the content remains inaccessible to treatment machines only human being able to interpret it. Semantic web has then ambition to remove this difficulty by combining the resources web ontological entities as semantic references, which allow different software to enter and directly exploit the content of the resource agents and reason above.

An example of a tag that would be used in a non-

semantic web page: <h1>Master Computer</h1>

Encoding similar information in a semantic web page

might look like this:

<h1 rdf:about="http://fstbm.ma.ac/informatique/ masterComputer">Master Computer</h1>

Collecting, structuring and recovery work related data

using technologies that provide a formal description of concepts, terms and relationships within a given knowledge domain. Some technologies already specified and standardized on which to build the Semantic Web (Fig 1):

XML (Extensible  Mark-­‐up  Language) provides a basic

syntax for content structure within documents, but it does not describe the semantics of the document. XML is not at present a necessary component of the Semantic Web technologies in most cases, as there are alternative syntaxes Turtle.

Turtle (Terse   RDF   Triple   Language) is a de facto standard because it is less verbose than XML, but has not been selected through a formal standardization process.

XSD (XML Schema Definition) is a language for describing XML document format for defining the structure and content type of an XML document. This definition enables you to check the validity of this document.

RDF (Resource Description Framework) is a simple language for expressing data models as objects ("resources") and their relationships. A model based on RDF can be represented through several syntaxes exchanges, RDF / XML, Turtle, and RDFa. RDF is a fundamental norm of the Semantic Web.

RDFS (RDF Schema) extends RDF and vocabulary to structure properties and classes in a resource described in RDF.

OWL (Web Ontology Language/ Ontology Vocabulary in (Fig 1)) adds more vocabulary for describing properties and classes: as with relations between classes, cardinality ("exactly one"), equality, the typing of properties, characteristics of properties (symmetry), etc..

Fig 1 Semantic Web Stack  

SPARQL (SPARQL Protocol and RDF Query Language  pronounced, "sparkle") is a query language and protocol that allow you to search, add, modify or delete RDF data available on the web through Internet.

URI (Uniform Resource Identifiers) are a means of identifying resources, especially but not exclusively on the Web. URIs have two main uses, and one URI can be used in both senses:

• Uniform resource names (URNs) are URIs that are used to name something, even if this is an abstract object that is not available on the Web. For example, a person can have a URI that is used in ontologies to refer to that person.

• Uniform Resource Locators (URLs) are URIs that are used to specify the location of something. URIs must start with a protocol identifier, and URLs typically use protocols that have a well-established technical interpretation that tools can use to access the specified location ("http").

Trust layer ranges from digital signatures and security to social network analysis.

Proof layer supports the exchange of “proofs” in an inter-agent communication enabling the common understanding of how the desired information is derived.

Logic layer enables intelligent reasoning with meaningful data.

 2. EVOLUTION OF ONTOLOGY

According to Stojanovic [2], the evolution of an

ontology is adapting to change and consistent propagation of these changes to dependent artefacts ( referenced objects, dependent ontologies and applications software using the ontology).

Currently, there are basically two approaches that deal with the evolution of the ontology: the first approach stojanovic [2], Giorgos [3] is based on the creation of a new ontology to adapt to the changes involved. In this approach, a single version of the ontology is maintained; whichever is later. To achieve the development of ontology, an evolutionary process is used. It consists of six main stages (fig 2): detection, representation, semantics, implementation, propagation, and change validation.

Page 3: Managing Semantic Web evolution in the Organization (SWO)tal.ircam.ma/conference/data/papers/7.pdf · view, the evolution of ontology is a following instances of ontology evolution

Fig 2 The life cycle of Evolution of an ontology [4]

The essential role of the semantics of change phase

in the ontology evolution process is to figure out which elementary changes need to be performed for one change request. For example deletion of a concept. If this were left to an ontology engineer, the evolution process would be too error-prone and time consuming; it is unrealistic to expect that humans will be able to comprehend entire ontology and interdependencies in it.

There are many ways to achieve consistency after a change request. For example, when a concept from the middle of the hierarchy is being deleted, all sub-concepts may either be deleted, or reconnected to other concepts. If sub-concepts are preserved, then properties of the deleted concept may be propagated, its instances distributed, etc. Different ways for resolving the request for the removal of the concept “Student” by considering only the concept hierarchy are shown in (Fig 3).

The first solution (Fig 4) does not contain any sub-concepts of the concept “Student”.

The second solution (Fig 5), all sub-concepts are retained and connected to the concept “Person”, which is the parent concept of the concept “Student”.

The last solution (Fig 6) also preserves all sub-concepts while reconnecting them to the root of the concept hierarchy.

In this approach, researchers are focusing their efforts on the development of the evolutionary process and do not take into account the management of multiple versions of the ontology evolved.

Fig 3 The given ontology after applying only the removal of the concept “Student”  

Fig 4 The updated ontology where all sub-­‐concepts are

deleted  

Fig 5 The updated ontology where all sub-­‐concepts are reconnected to the parent concept  

Fig 6 The updated ontology where all sub-­‐concepts are reconnected to the root concept  

 However, the management of multiple versions is

required by distributed systems using ontology-evolved nature. So it is very difficult (we can say impossible) to apply or synchronize updates on the same version of the ontology according to Klein [5].

The second approach Klein [5] and Heiner [6] uses the term 'Versioning’, consists in creating and managing different versions of ontology. It addresses the problem of (partial) incompatibility of new ontology versions with the previous one and thus, with ontology’s instances and, applications and dependant ontologies. The evolution of ontology is defined as the ability to manage changes in the ontology and their effects in creating and maintaining different versions of ontology. This ability is to identify, differentiate and modify versions, specify relationships that make explicit the changes made between versions and use access mechanisms for dependent artefacts.

Page 4: Managing Semantic Web evolution in the Organization (SWO)tal.ircam.ma/conference/data/papers/7.pdf · view, the evolution of ontology is a following instances of ontology evolution

The authors do not propose an approach for supporting the evolution of the process ontology, but does not support versioning of ontology after its evolution. They then provide an analysis of the relationship between the model versions of the ontology, but without worrying about managing access to dependent artefacts (referenced objects, ontologies, applications) using versions of the ontology. In addition, the authors do not develop functional framework for integrating all elements of methodology they propose.

3. TOOLS SUPPORTING THE DEVELOPMENT OF ONTOLOGY

Manually manage changes in an ontology is a complex

and expensive task. The ontology engineer must control all the effects of change, solve and evaluate the impact of change on the ontology. There needs appropriate tools providing the technical means necessary for the evolution of ontology.

The evolution of ontology must be part of the functionality of an ontology editor to guide the ontology development in an iterative and dynamic process. Yet all existing ontology editors do not consider the changing needs of ontology. These needs are mainly [2]: functional needs, customization needs, the needs of transparency, the need for reversibility, audit requirements, needs refinement and usability needs.

Also some existing ontology editors support some aspects of evolution, current ontology development work - including some described in the previous sections - offer specialized tools whose objective is either to guide the user to manually apply changes is to manage and automatically apply the changes. Some of these tools allow collaborative editions; others support the aspects of ontology versioning.

3.1. ECCO EDITOR    

ECCO [8] is an editor with a collaborative and contextual environment ontologies. It covers the various stages of the development cycle of an ontology starting from the recovery of terms extracted by tools of word processing (NLP) to create a vocabulary that is then enriched, hierarchical and formalized in an OWL-Lite ontology.

It keeps the history of the process of developing ontology as metadata stored in RDF format and also can generate traces changes to the ontology.

3.2. KAON TOOLS    

The Framework KAON proposed by the University of Karlsruhe, implements an ontology evolution system (Fig 2). In addition to automating the process of evolution, the box KAON tools guide the ontology engineer to formulate its changing needs by providing supplemented by further information and suggestions for improvement of the ontology. It offers identification functionality needs change data driven, allows users to define "evolution strategies" monitoring the implementation of changes, and facilitates the

adaptation of the ontology to the needs of end users (identification of needs-led use).

3.3. Protégé EDITOR  

 Protégé is a tool developed by the Medical Informatics

Group at Stanford. It includes an ontology editor, open source, and a Framework knowledge base. Is protected with a graphic and interactive design environment ontology and knowledge acquisition. Its architecture is component-oriented facilitating the enrichment of editor features by adding plugins (such as Protégé-OWL plugin). Some plug-ins are specific aspects of ontology evolution and are presented in the sections below.

3.4. PROMPTdiff TOOL  

 PROMPTdiff a plug-in developed for Protected for

finding mappings between frames based on heuristics- have been proposed (Klein). Their role is to define the evolutionary relationships between the elements of two versions of an ontology. User interface allows you to view some complex changes between versions of ontology.

3.5. Plib Editor  

 This editor [9] is based on an architecture database called

to ontological basis (OntoDB), which allows storing in the database, both the ontology model and ontology instances.

4. SEMANTIC WEB ORGANIZATION APPROACH

This work is the realization of an editor that supports all

the problems of the evolution of ontology, preferment more than existing tools. I have based my ideas on these "Managing the evolution of the Semantic Web Company [4]" (Fig 7). to overcome the problems the evolution of ontology, we ses three human agent in this system :  

• Ontology Provider: provides the ontology used and changes in the system.

• Annotator: creates and modifies the annotation using ontology provided.

• System Engineer: manage system functions. Such agents interact with the system through the

notifications; the ontology provider sends a message of creation or change of ontology to the system engineer. The system engineer applying inconsistency detection rules to search annotations, which become obsolete and inconsistent with respect to the new version of ontology. Annotator annotates the resources of the company based on the concepts and relationships defined in the former ontology.

Several problems we correct it among them, Jena API (A) for storage of ontology in database to use all the power of a DBMS (B), more implementer a system for reasoning in many instances eliminate all inconsistencies by FaCT++ or Pellet.

Page 5: Managing Semantic Web evolution in the Organization (SWO)tal.ircam.ma/conference/data/papers/7.pdf · view, the evolution of ontology is a following instances of ontology evolution

Fig 7 Architecture of a Corporate Semantic Web  [4]  

 III. CONCLUSION

The work presented in this manuscript does not purport to be a complete solution to all aspects of the problem of evolution of the Semantic Web organization. There are still some limitations and problems in the system as there are no any tools or method addresses how to automatically connect concept, and how each new properties to our knowledge and it would be unrealistic to determine automatically a choice of strategies resolution for these types of complex changes. References [1] V. Sugumaran et J. A. Gulla, Applied Semantic  

Web Technologies. CRC Press, 2012.  [2] M.Sc. Ljiljana Stojanovic, « Methods and Tools for Ontology

Evolution », Ph.D, Fridericiana zu Karlsruhe genehmigte Dissertation., 2004.  

[3] G. Flouris et D. Plexousakis, « Handling Ontology Change: Survey and Proposal for a Future Research Direction », 2005.  

[4]    P.  H.  Luong,  «  Gestion  de  l’évolution  d’un  Web  sémantique  d’entreprise   »,   École   Nationale   Supérieure   des   Mines   de  Paris,  2007.  

[5]    N.  F.  Noy  et  M.  Klein,  «  Ontology  Evolution:  Not  the  Same  as  Schema  Evolution  »,  Knowl.  Inf.  Syst.,  vol.  6,  no  4,  p.  

428�440,  juill.  2004.  [6]     H.   Stuckenschmidt   et   M.   Klein,   «   Reasoning   and   change  

management  in  modular  ontologies  »,  Data  Knowl.  Eng.,  vol.  63,  no  2,  p.  200�223,  nov.  2007.  [7]    H.  Stuckenschmidt,  L.  Serafini,  et  H.  Wache,  «  H.:  Reasoning  

about  ontology  mappings  »,  2005.  [8]     «   e-­‐WOK   HUB :   Résultats   -­‐   ECCO   ».     http://www-­‐

sop.inria.fr/edelweiss/projects/ewok/publications/ecco.html  

 [9]    «  PLIB  »http://www.plib.ensma.fr

Page 6: Managing Semantic Web evolution in the Organization (SWO)tal.ircam.ma/conference/data/papers/7.pdf · view, the evolution of ontology is a following instances of ontology evolution