Managing semantic annotations evolution in the CoSWEM system

Managing semantic annotations evolution in the CoSWEM system Luong Phuc Hiep 1, 2, Rose Dieng-Kuntz 1, Alain Boucher 2 1 INRIA Sophia Antipolis, France 2004 route des Lucioles, BP 93, 06902 Sophia Antipolis, France. (Phuc-Hiep.Luong Rose.Dieng@sophia.inria.fr) 2 Institut de la Francophonie pour l'informatique, equipe MSI, Vietnam Bâtiment D, 42 rue Ta Quang Buu, Hanoi, Vietnam. (lphiep aboucher@ifi.edu.vn) Abstract. Knowledge is considered as one of the most important assets of organisations, which decisively influences its competitiveness. An approach for managing knowledge in an organisation consists of building a Corporate Semantic Web (CSW). The main components of a CSW are (i) evolving resources distributed over an intranet and indexed using (ii) semantic annotations expressed with the vocabulary provided by (iii) shared ontologies. However, the CSW components are evolving and they need to be managed in the organisation knowledge management system. In this paper, we study a case of modification in the CSW. Changes in the ontology may impact to the consistency of semantic annotations that use concepts or properties defined in this ontology. We present the CoSWEM (Corporate Semantic Web Evolution Management) system enabling to manage the evolution of such a CSW, especially to address the evolution of semantic annotations by applying a rule-based approach allowing to detect and to correct semantic annotation inconsistencies during its evolution. Keywords. Semantic Web, Corporate Semantic Web, Ontology, Semantic Annotation, Evolution. 1. Introduction Organisation's knowledge assets concern their resources, business and management experiences, services, applications and also the resulting lessons learned. These knowledge assets become the competitiveness and survival for many organisations since they have to react to changes frequently occurred in their domain and activities. In this dynamically changing context, building up a corporate memory [10] is necessary for succeed of organisation that can meet its primary organisational objectives: improving knowledge sharing, cooperative work and relationships with the external world, reuse of knowledge [12]. In the next generation Semantic Web aiming at a better cooperation among humans and machines, corporate memory can be materialised as a Corporate Semantic Web (CSW), which composed of heterogeneous, evolving resources distributed over an intranet and indexed using semantic annotations expressed with the vocabulary provided by shared ontologies. However, organisations find themselves in dynamic and changing environments because of the changes in their business, technologies and processes These changes in the real world often result in need of modifications of the CSW. For example changes in business environment lead to continuous changes in the conceptualisation of a business domain that might impact on the underlying domain ontologies. These changes in underlying ontologies then need to be propagated to all semantic annotations that are created based on the shared vocabulary of ontologies. Consequently, such modifications influence the activities and performances of the CSW system. In this paper we present the CoSWEM system enabling to manage the evolution of such a CSW when its ontologies or the semantic annotations of its resources change. This

proposed system will support to the propagation of the ontology changes towards the semantic annotations depending on this ontology by providing a rule_based approach to solve the annotation inconsistencies. First of all, we present the background work with the introduction of some terminologies related to our work in section 2. Then, section 3 describes a scenario of ontology modification and the way to solve this modification in the CoSWEM system. Finally, we make a survey of current related researches in section 4 before giving the conclusion. 2. Background work 2.1 Semantic Web and related technologies The development of the actual Web is fast with a huge number of users and the amount of information. However, most of the information that is available has to be interpreted by humans but machine support is rather limited. In order to get rid of that limitation, Tim Berners-Lee presented the vision of the Semantic Web to make the contents of the actual Web accessible and interpretable by machines [14]. For a better working cooperation between computers and people, information is given well-defined meaning provided by ontologies. These ontologies are used to annotate the domain resources as machineprocessable data called semantic annotations. Ontologies provide consensus semantics of data and information sources that can be communicated between different agents (software and people). Authors of [13] defines an ontology as a formal explicit specification of a shared conceptualization. An ontology comprises a set of concepts describing a domain and a set of relationships between these concepts. The Semantic Web proposes annotating document content using semantic information from domain ontologies. Semantic annotations provide a more precise description of the knowledge contained in the document and its semantics in the domain. They contain information that is made understandable for the computer, thus assisting people to search, extract, interpret and process information. 2.2 Corporate Semantic Web approach With the purpose of deployment organisational knowledge using Internet and web technologies for an organisation, many researches have been carried out to study the integration of the corporate memory in a new environment Semantic Web that aims at making the semantic contents of Web resources understandable not only by human but also by programs [14]. The research team ACACIA 1 has been studying the materialization of a corporate memory as a Corporate Semantic Web [12] consisting of: - Resources: they can be documents, databases (represented in various data formats such as XML, HTML or even classic formats). These resources also correspond to people, services, software or programs. - Ontologies: they describe the conceptual vocabulary shared by one or several communities in the organisation 1 Research team at INRIA Sophia Antipolis (http://www-sop.inria.fr/acacia/) is working towards offering methodological and software support (i.e. models, methods and tools) for knowledge management (i.e. for building, managing, distributing and evaluating a corporate memory) for an organisation or community.

- Semantic annotations: they use the conceptual vocabulary defined in ontologies to describe resources, for example contents of documents, skills of persons or characteristics of services, programs or software Fig. 1 - Architecture of a Corporate Semantic Web The architecture of a CSW presented in the Fig. 1 consists of three essential components resources, ontologies and semantic annotations base. All the semantic annotations that describe resources by using the shared vocabulary of ontologies are stored in the knowledge base. Individual (collective) users can browse useful knowledge with the help of several tools and semantic search engine that are integrated in the knowledge management system. 2.3 Evolution problems in a Corporate Semantic Web When one of three main components of a CSW is changed, it might impact to the consistency in one part or in overall system. In this case, others related parts may need to be evolved as well to meet the changes. Among these three components, ontologies are the evolving factor since their domains are changing fast (new concepts evolve, concepts change their meaning, new business rules are defined, etc.). [7] has proposed three types of changes in ontology evolution: changes in domain, changes in conceptualisation and changes in specification. A modification in one part of the ontology may generate inconsistencies in other parts of the same ontology, in the ontology-based instances as well as in depending ontologies and semantic annotations [6]. Semantic annotations are less evolved because they are used only for the description of the content of knowledge sources. Like the ontologies, resources also influence on the semantic annotations in case of change. However, within the scope of research, we have been concentrating on the influences of modified ontologies on their semantic annotations expressed with the vocabulary provided by these underlying ontologies. 3. Semantic annotation evolution in the CoSWEM system A modification in one part of the ontology may generate inconsistencies in other parts of the same ontology, in the ontology-based instances as well as in depending ontologies and in the applications using this modified ontology [5]. Moreover, changes in ontology may impact to the semantic annotations that use concepts or properties defined in this ontology. In this section, we emphasize the propagation of the ontological changes to the semantic annotations in order to keep an overall consistency. This change propagation phase can be considered in the process of semantic annotation evolution.

3.1 Scenario: semantic annotation evolution when underlying ontology changes We examine a scenario in which a biologist (or a doctor) needs to create semantic annotations describing a patient s profile. These semantic annotations use several specialized terms that are defined in the biomedical UMLS 2 ontology containing the concept Flu and its sub-concept Bird_Flu. After some executed changes in the biomedical UMLS ontology (e.g. the concept Bird_Flu is deleted), some existing statements of annotation may be inconsistent with respect to the ontology because of lack of the reference to the deleted concept Bird_Flu. One possible solution in this case is to replace the deleted concept name Bird_Flu in annotation by its parent s name Flu. Ontology before removing the concept <ev:person rdf:about="http://info.com/j.ben"> <ev:hasdisease rdf:resource= &ev;bird_flu > </ev:hasdisease> </ev:person> Annotation before updating statement Ontology after removing the concept <ev:person rdf:about="http://info.com/j.ben"> <ev:hasdisease rdf:resource= &ev;flu > </ev:hasdisease> </ev:person> Annotation after updating statement 3.2 Process of semantic annotation evolution We present a process enabling to address the inconsistency of semantic annotations. We choose RDF(S) languages [15], which are recommended by W3C, to formalize our ontologies and annotations. The process includes the following main steps: Step 1 : convert UMLS metathesaurus and semantic type to ontology in RDF(S) The UMLS metathesaurus and semantic types containing information about biomedical which are represented in structured textual files, so they need to be converted into a suitable format of an ontology in RDF(S) for later processing. We also convert the log file capturing realised changes between two metathesaurus versions to appropriate format (e.g. this change log format can include some information related to the author of change, type of change, the old/new value, etc.). This converted log file allows determining quickly the annotations related to ontology changes. In our project, we have built an ontology in language RDFS in order to describe the biological domain (drugs, genes, processes). This ontology consists of concepts and relations extracted from UMLS semantic type network and a part of metathesaurus [4]. Step 2 : query semantic annotations, using Corese: 2 Unified Medical Language System (http://www.nlm.nih.gov/research/umls/meta2.html)

We use Corese 3, an ontology-based search engine for the semantic web [9], that is dedicated to the retrieval of web resources annotated in RDF(S). Corese proposes a query language for RDF very similar to SPARQL 4 and provides a rule language enabling us to deduce new knowledge from existing one. It makes inferences on the whole annotation according to user s queries. In this step, Corese especially can be used to retrieve from existing annotation bases: - annotations related to modified ontology by using their reference link to this ontology. - potential inconsistent annotations (they may include both related consistent and inconsistent annotations) by using a set of ontology changes. A potential inconsistent annotation means that it relates to the ontological change but its consistency constraint has not been verified. Step 3 : apply inconsistency detection rules We apply inconsistency detection rules in this step in order to detect the real inconsistent annotations from a set of potential inconsistent annotations. A real inconsistent annotation means that it violates the consistency constraint defined for the annotation. These detection rules are formulated from a set of constraints and are expressed in the syntax of Corese rule language. Step 4 : apply inconsistency correction rules and resolution procedures After having determined real inconsistent annotations from a set of potential inconsistent annotations, these annotations will be repaired by applying correction rules. We have established all possible solutions that solve the propagation of ontological changes to their semantic annotations in order to keep consistency status. Step 5 : versioning management This step enables us to manage versions of the ontology and its semantic annotation in case of storing different versions. 3 COnceptual REsource Search Engine (http://www-sop.inria.fr/acacia/soft/corese/) 4 http://www.w3.org/tr/rdf-sparql-query/

Fig 2- Process of semantic annotation evolution The actors involved in this system (as Fig.2) are the System_Engineer who manages the overall system, the Annotator who creates annotations based on existing ontology, the Ontology_Provider who provides changing ontology source. 3.3 Rule-based approach for solving inconsistencies of semantic annotation The rule-based approach is constructed from some consistency constraints that must be satisfied for annotation model. Therefore, we have proposed some consistency constraints that can be considered as an agreement among semantic annotations entities with respect to their underlying ontology. Here below is an example of a consistency constraint on concept: Constraint on concept: All the concepts used in the annotation must be defined before in the ontology. Based on the consistency constraint, we have established several inconsistency detection rules enabling us to find inconsistent annotations that are obsolete related to a modified ontology. These rules are based on consistency constraints and are represented in syntax of Corese rule language (including some primitives in SPARQL). For example, a rule that detects inconsistent annotations whose concept has not been defined before in the ontology can be declared as the following: Detection rule for concept resource: If a concept is used in an annotation but it has not been defined in the ontology, then this annotation leads to inconsistent status. This detection rule can be We use the syntax of Corese rule language to represent expressed as following: this detection rule as below: <cos:rule> <cos:if>?x rdf:type?c IF option(?c rdf:type?class?x rdf:type?c?class = rdfs:class not(?c rdf:type rdfs:class) ) THEN filter! bound(?class) error("inconsistent") </cos:if> <cos:then>?x rdf:type cos:error </cos:then> </cos:rule> After having detected and collected all inconsistent annotations from a set of potential inconsistent annotations, we need to correct these inconsistencies by applying some correction rules on these annotations (e.g. what should the system do if a leaf concept is deleted?). We have established several resolution procedures for each ontological change (e.g. how to solve an inconsistent annotation if it relates to a deleted leaf concept in ontology?) and specified how to propagate the change resolution to inconsistent annotations in order to keep an overall consistency. We examine a simple example of the deletion of a leaf concept in the ontology. Suppose we have a leaf concept Bird_Flu which is the child of an super-concept Flu, a property p that may receive Bird_Flu or Flu as its domain/range. Now we make a change in the

ontology by removing the leaf concept Bird_Flu. After having deleted this leaf concept in the ontology, several statements in annotations related to this concept Bird_Flu may be in inconsistent status with respect to this modified ontology. To solve this problem, there are several possible solutions (see table below) that can be applied to obsolete annotations depending on each ontological change resolution. Inconsistency correction rules If Bird_Flu and Flu are the domains/ranges of p in ontology, Then system will replace the ressource (domain/range of p) in annotation which is instance of the concept Bird_Flu by its super-concept Flu. If Bird_Flu is the domain/range of p and Flu is not the domain/range of p in ontology, Then system will delete all statements including the instance of concept Bird_Flu in annotation. Resolution Procedure All obsolete statements including the instance of concept Bird_Flu (which is the domain/range of p) will be replaced by new statements including the instance of its super-concept Flu. All obsolete statements including the instance of concept Bird_Flu will be removed. 4. Related works The authors in [3] point out that ontologies on the Web will need to evolve but they does not mention to the change propagation between distributed and dependent ontologies. There are also some studies of managing changes in ontologies [8] [6] that take ideas from research on database schema evolution [2]. Paper [8] describes an ontology versioning system that allows access to data through different versions of the ontology. This approach is based on the comparison of two ontology versions in order to detect changes. In contrast to that approach of versioning, which detects changes by comparing ontologies, the authors of [6] present their ontology evolution approach that allows access to all data only through the newest ontology. The authors of [1] have introduced a combined approach that supports the process of ontology evolution and versioning by managing the history of ontology changes. However, all these approaches mentioned above have not treated the change propagation to the related annotations in case of changes in ontologies. There are very few approaches investigating the problems of propagation of the ontological changes to semantic annotations. The authors of [7] proposed a framework CREAM to solve the evolution of metadata based on their existing research on the ontology evolution. However, this approach only presented their proposition of a framework for enabling consistency of the descriptions of the knowledge sources in case of changes in the domain ontology but they do not specify techniques to solve it. Another study of ontology evolution influence on metadata via relational constraints of a database system is also given in [11]. According to this approach, the knowledge-based environments rely on a relational database to store the RDF and RDFS used for representing respectively ontology-based assertions and the ontology structure itself. Ontology maintenance events can be managed using database triggers for automatically modifying property ranges or domains in the stored assertions.

5. Discussion and Conclusion The work presented in this paper can be compared with some similar existing studies. Regarding to the solving inconsistencies in semantic annotations in the context of modifications in ontology, authors in [7] has presented the framework CREAM enabling the consistency of semantic annotations when ontologies changes but have not specified a particular technique to solve it. The technique of using triggers in relational database for automatically modifying property range/domain in the stored assertions has been introduced [11] but it has not mentioned the inconsistency resolution process. Our approach not only proposes the propagation process of the ontological changes to semantic annotations but also specify a rule-based approach to detect inconsistent annotations and the correction procedures to solve these inconsistencies. We have presented in this paper the CoSWEM system enabling to manage semantic annotation evolution in the new infrastructure of Semantic Web. The evolution process and the rule-based approach have addressed inconsistent effects on annotations when the ontologies change. As further work, we will study some effective algorithms on the process of correction and validation for semantic annotations changes. We will also focus on the problem of versioning management allowing to use different versions of semantic annotations and ontologies. References [1] D.Rogozan and G.Paquette (2005). Managing Ontology Changes on the Semantic Web. Proceedings of the 2005 IEEE/WIC/ACM International Conference on Web Intelligence. [2] J.F.Roddick. A Survey of Schema Versioning Issues for Database Systems, Information and Software Technology, 37(7):383-393, 1996. [3] J.Heflin, JA.Hendler (2000). Dynamic ontologies on theweb. Proceedings of the 7 th national conference on artificial intelligence AAAI-2000, Menlo Park, CA, August 2000, pp 443449. AAAI/MIT Press, Cambridge, MA [4] K.Khelif, R.Dieng-Kuntz, P.Barbry (2005). Semantic web technologies for interpreting DNA microarray analyses: the MEAT system. Proc. of the 6th International Conference on Web Information Systems Engineering WISE 05, New York, 20-22 novembre 2005. [5] L.Stojanovic (2004). Methods and Tools for Ontology Evolution. PhD thesis, University of Karlsruhe, 2004. [6] L.Stojanovic, A.Maedche, N.Stojanovic and B.Motik (2002) User-driven ontology evolution management. In European Conf. Knowledge Eng. and Management (EKAW 2002), pages 285 300. Springer-Verlag, 2002. [7] L.Stojanovic, N.Stojanovic and S.Handschuh (2002). Evolution of the Metadata in the Ontology-based Knowledge Management Systems, Proceedings of the 1st German Workshop on Experience Management: Sharing Experiences about the Sharing of Experience, 2002. [8] M.Klein (2004). Change Management for Distributed Ontologies. PhD thesis, Vrije Uniersiteit Amsterdam, Aug. 2004. [9] O.Corby, R.Dieng-Kuntz, C.Faron-Zucker (2004). Querying the Semantic Web with the CORESE search engine. Proc. of the 16th European Conference on Artificial Intelligence (ECAI 2004), Valencia, 22-27 August 2004, IOS Press, p. 705-709 [10] O.Kuhn and A.Abecker (1997) Corporate memories for Knowledge Management in Industrial Practice: Prospects and Challenges, Journal of Universal Computer Science, vol. 3, no. 8. [11] P.Ceravolo, A.Corallo, G.Elia and A.Zilli Managing Ontology Evolution Via Relational Constraints. Knowledge-Based Intelligent Information and Engineering Systems, 8th International Conference, KES 2004, Wellington, New Zealand. Lecture Notes in Computer Science 3215 Springer 2004.

[12] R.Dieng-Kuntz, O.Corby, F.Gandon, J.Golebiowska (2004). Gestion dynamique des connaissances industrielles. Chapitre 1. Lavoisier, 2004. [13] R.Gruber. A translation approach to portable ontology specifications, Knowledge Acquisition, An International Journal of Knowledge Acquisition for Knowledge-Based Systems, Volume 5, Number 2, pp.199-220, 1993. [14] T.Berners-Lee, J.Hendler and O.Lassila (2001) The Semantic Web, In Scientific American, May 2001, p35-43 [15] W3C Resource Description Framework, http://www.w3.org/rdf.