Search

Dit document is een verslag van de werkzaamheden in de vierde periode van het PoCeHRMOM project (IWT-TETRA-50115). Deze periode ving aan op 1 september 2006 en eindigt op 31 december 2006. We schetsen de huidige status van het project en letten hierbij in het bijzonder op de technische (sectie 1), beheers- (sectie 2) en valorisatieaspecten (sectie 3).

Hand-out PoCeHRMOM vergadering 09/11/2006

Op het Semantisch Web worden websites geannoteerd volgens overeengekomen metadata via standaarden zoals het Resource Description Framework (RDF). Semantische metadata laten webapplicaties toe om informatie uit diverse sites op betekenisvolle manier te manipuleren en uit te wisselen. Het inhoudelijk structureren van sites gebeurt via ontologieën. Een ontologie is een formeel gestandaardiseerd classificatiesysteem waarin wordt aangegeven welke betekenisvolle elementen of concepten er binnen een bepaald kennisdomein bestaan en hoe deze aan elkaar worden gerelateerd. Onderzoek in het Europese FF POIROT project (IST 2001-38248) heeft ondermeer aangetoond dat terminologiebeschrijvingen, voorzien van rijke semantische informatie, bijdragen tot een efficiënte en kwaliteitsvolle opbouw van domein gerelateerde ontologieën. Dit doctoraatsonderzoek bouwt verder op deze vaststelling en wil onderzoeken hoe betekenisverschillen tussen termen, als gevolg van applicatieve en culturele betekenisvorming, via linguïstische en terminologische benaderingen kunnen geïntegreerd worden in ontologieën. Het project omvat in een eerste fase een literatuurstudie van bestaande methodes voor het beschrijven van betekenisvariatie. Vervolgens worden de theorieën aan de praktijk getoetst d.m.v. een aantal specifieke casusstudies.

Within the framework of the FF POIROT project, the Centrum voor Vaktaal en Communicatie has worked out the Termontography methodology for developing (multilingual) ontology-based terminological resources. In an ontology-based terminological resource, the terminology extracted from texts is structured according to a predefined set of categories and relations, which scope both the information gathering as well as the search phases in the Termontography method (Kerremans et al. 2003). In an article presented at the EURALEX international conference, CVC noted that in the commercial terminology management systems we know of, the organization of multilingual terminology in a network of categories (also known as categorisation framework) is not possible (Kerremans et al. 2004). Although these systems are concept-oriented (i.e. multilingual terminology denoting the same concept is placed in the same terminological record), they do not allow users to specify how the concepts are related within a domain of interest. Therefore, a terminological resource developed in these systems hardly has an ontological foundation. In Kerremans et al. (2004) CVC wrote down some important user requirements for a workbench supporting the development of (multilingual) ontology-based terminological resources. A programmer was recruited in the last year of the FF POIROT project to develop a software environment that meets these requirements. As this development phase is still in progress, this report will discuss the intermediate results and present the work achieved so far. This document is structured as follows: section 2 summarises our user requirements. Section 3 covers the technical description of the software workbench. In particular, we focus on the implementation of the Termontography workbench that allows users to develop ontology-based terminological resources (section 3.1). In section 3.2, we introduce the KBExplorer tool that allows us to query and search through such resources. Section 4 states the Intellectual Property Rights. In section 5, we conclude.

Due to its huge volume of information and wide-spread access, the Internet becomes an attractive medium for fraudsters to reach a large number of potential victims in a short time. Moreover, its size and dynamics make it difficult to put the web contents under close legal surveillance. Consequently, most frauds on the Internet are discovered by accident and a large number of them go undetected (Davia, 2000). One of the aims of the FF POIROT project (IST-2001-38248), a European research project in the fifth framework is to develop formal and shareable knowledge repositories (i.e. ontologies) and terminological resources for applications detecting and intercepting e.g. securities fraud on the Internet. Securities fraud refers to the selling of overpriced or worthless shares, or other financial instuments to the general public (Zhao et al., 2004c). Within the framework of this research project, two approaches have been worked out to support the development of ontological repositories, on the one hand, and terminology bases, on the other hand. The integration of these resources in fraud detection systems, motivates the application-oriented views in both these methodologies: AKEM (for knowledge engineering) and Termontography (for terminology engineering). This article wants to show how the methods of AKEM and Termontography are interacting in the development process towards an innovative and technological solution for detecting and intercepting securities fraud. The development tracks will be outlined in section 2. Next, we will discuss the Termontography (section 3) and AKEM (section 4) approaches and explain how both approaches interact within the scope of the problem and knowledge space (section 5). In particular, we will show how principles of the AKEM methodology are integrated in Termontography and, vice versa, how Termontography contributes to the development of the formal and shareable knowledge repository. Finally, in section 6 we conclude.

Dit is een verslag van de PoCeHRMOM vergadering van 07.09.05

The VAT topical ontology is a formal knowledge resource developed by CVC and VUB. The content of the VAT topical ontology roughly coincides with the information presented on the website of the European Taxation and Customs Union . The structure of this website has determined the general scope of the ontology whereas information about the categories or topics discussed in the webpages are constructed by applying the AKEM knowledge engineering approach. AKEM is based on the DOGMA modelling approach (Jarrar and Meersman 2002) and primarily emphasises a bottom-up (i.e. text-driven) application-oriented view in ontology modelling (Zhao 2004; Zhao et al. 2004). To some extent, the methodology was also found useful for the development of a domain-specific ontology which surpasses the boundaries of a univocal application-oriented view, provided that the ontology is seen as a knowledge repository serving more than one application scenario in the domain of interest. This report not only discusses the VAT topical ontology created in the FF POIROT project. It also focuses on Termontography, the approach in terminography which has been worked out by CVC Brussels (Kerremans et al. 2003; Kerremans 2004; Temmerman and Kerremans 2003).

This document discusses a number of experiments that were set up in order to calculate and compare the number of terms and roles extracted by CVC and STAR Lab VUB from the European Sixth Directive on VAT (77/388/EEC). Next we present the Applied Knowledge Engineering Methodology (AKEM) in which both terminographers and ontology engineers collaborate in the development process of multilingual ontologies. The report shows how CVC has been able to isolate from texts certain knowledge patterns that are important in this development process.

This document discusses the content of the VAT topical ontology created in the FF POIROT project. The ontology will provide the concepts and relations used in several VAT applications for document management, monitoring, question-answering, etc. The scope of the VAT ontology is to a large extent based on the structure of the website of the Taxation and Customs Union of the European Union.

This document summarises the contribution of CVC to WP8 in the FF POIROT project.

Terminology management is part and parcel of many knowledge applications. Due to their availability in electronic formats, terminological databases are now used in a way that surpasses their traditional role as terminological dictionaries for human users. In this paper we explain how the method of termontography, developed at CVC Brussels, can be used in the development of specialised terminological resources. Termontography is a multidisciplinary approach in which theories and methods for multilingual terminological analysis (based on sociocognitive terminology theory) are combined with methods and guidelines for ontology engineering. A clear distinction is made between conceptual modelling at a language-independent level and a language and culture specific analysis and description of discourse-related units of understanding. Some critical issues are raised concerning the feasibility and limits of next generation terminological resources. Examples will be given of how the limitations ofparticular tools for taxonomy and ontology building (e.g. Protégé, Ontoterm) in combination with (multilingual and multicultural) terminology description force us into questioning some of the basics of terminology theory and knowledge representation theory. We will raise some issues on the evolution and dynamics of meaning and understanding starting from terminological records i.e. units of description in terminology management and moving towards flexible and multidimensional units of understanding. The theoretical considerations are supported by examples from case studies that were executed in several projects at CVC Brussels.

Our research in the FF POIROT project has shown that ontologies can be used to extract the semantics of fraud evidence. Starting from a conceptualised model of Nigerian fraud e-mails results have indicated that an information extraction application, can recognise and collect so-called red flags or fraud indicators in e-mails. In this document we discuss how the information extraction application analyses French Nigerian fraud letters. The idea behind all this is to come up with an application that is able to automatically intercept Nigerian fraud letters and sends these to the regulatory authorities in each country.

Given the general architecture of the VAT topical ontology created in the FF POIROT project, this document reflects on how applications can make use of this formal knowledge repository. In particular, this document looks at three possible applications in which the ontology can play a role. In section 2, we explain the idea of an ontology-based information extraction tool which assists both tax authorities as well as traders in extracting specific knowledge about companies. In section 3, we describe a possible document management system which allows users to perform concept-based queries in order to retrieve relevant legislative texts or even particular sections in VAT law. Finally, section 4 deals with a so-called reader-assistant tool which provides extra VAT-related information about categories which are mentioned with very few additional information in company websites.

This document summarises our progress in WP4 with respect to the development of a quadrilingual terminological database (French, Italian, Dutch and English) which is to deliver a multilingual layer to the financial fraud formal knowledge repository. We report on the semi-automatic extraction process of the multilingual VAT terminology derived from the VAT Sixth Directive corpus (77/388/EEC), a European directive dealing with the harmonisation of the European national laws on value added tax of which only the Italian, English, French and Dutch versions have been considered.

This document describes the progress, challenges and prospects of the FF POIROT project, covering the second year of the project implementation. Detailed information can be found in the other documents in the documentation package.

This document is the official FF POIROT deliverbale describing the multilingual terminological database developed by CVC. Section 1 deals with the current workflow we applied in order to build a large multilingual terminology repository in a short period of time. The second section discusses the sources that have been consulted for creating this large term base. In section 3, we summarise the fields used in the first release of the database and which fields have been removed from or added to the second release at the end of September 2004. Section 4 focuses on figures and data obtained from the first release of the terminology base at the end of March 2004. In section 5, we reflect on the practical value of the multilingual terminology base with respect to ontology development in WP2 and on its role as a computational utility in the fraud applications envisaged.

This document summarises the main activities and achievements of the Centrum voor Vaktaal en Communicatie (CVC) in the FFPOIROT project, starting from the 1st of June 2003 until the 29th of February 2004.

In this document, we present figures and data of the multilingual terminological database created by CVC in the FF POIROT project. The results are presented at the sixth consortium meeting in Edinburgh (3rd and 4th June 2004).

This document summarises the content of the multilingual terminological database created by CVC in the FF POIROT project.

Het Vlaams-Nederlands meerjarenprogramma voor Nederlandse taal- en spraaktechnologie STEVIN benadrukt ondermeer de nood aan een (eventueel modulair ingericht) multifunctioneel elektronisch lexicon ter ondersteuning van de annotatie van hieraan gekoppelde corpora. Dit lexicon dient niet alleen de algemene woordenschat af te dekken, waardoor het van belang is te starten van reeds bestaande lexica (bv. CELEX, RBN, Nederlandse EuroWordNet), maar met name voor eventuele domeinspecifieke corpora ook de relevante terminologie. Belangrijke behoeften omtrent de inhoud van een dergelijk lexicon werden reeds gedeeltelijk geformuleerd in de TST-basisvoorzieningen van het STEVIN programma. Zo dient ondermeer de relaties tussen ingangen in het lexicon en daadwerkelijke voorkomens van woordvormen in het corpus (incl. spellingvarianten, typefouten en tokenisation-fouten) behandeld te worden en is informatie omtrent bijvoorbeeld spelling, woordsoort, collocaties en semantische features essentieel voor het annoteren van corpora voor velerlei doeleinden.

In this report we show a first version of the meta-model of the FFPOIROT terminological database. This meta-model is necessary to be able to convert the FFPOIROT terminology base to an agreed, abstract lexon model which will eventually lead to a direct upload of the FFPOIROT terminology base into the FFPOIROT ontology server.

More than ever before terminology as a discipline is growing into multidisciplinarity. On the one hand, it is inspired by new developments in other disciplines like computational linguistics, artificial intelligence or database management. On the other hand, it can be at the service of other disciplines. The example we would like to elaborate on is how a multilingual terminographic analysis can contribute to the quality of the output of ontology engineers. The example we will use to illustrate this is taken from the FFPOIROT project , a European research project in the 5th framework which aims at compiling for several languages (Dutch-Belgium, French-France, Italian-Italy and English-UK) a computational knowledge repository (i.e. an ontology) for the financial forensics domain. The method of termontography was developed by CVC Brussels in order to add a multilingual layer to this ontology. This paper is structured as follows: in the first section we briefly explain some fundamental issues related to DOGMA modelling and show how our terminological resource is used by DOGMA ontology engineers in the FFPOIROT project (section 1). In the second section we elaborate on the problems that may arise when trying to match the content in our terminology resource with the ontology developed in DOGMA (section 2). In the third section we motivate the need for a framework of domain-specific categories that precedes the termontographic analysis. The fourth section summarises the termontography method and in section 5 we draw some general conclusions with respect to the integration of multilingual, termontological support in an ontology engineering framework.

In this document we report on the proceedings regarding the terminology and relation extraction from two case-studies in securities fraud: the Livingston and SmallXChange stories.

We focus on the unit-of-understanding approach Termontography (Kerremans et al. 2003; Temmerman and Kerremans 2003) and on the requirements concerning a workbench supporting this approach. The main concern is to facilitate the creation of (multilingual) domain-specific dictionaries that hold information such as how a term or phrase is related to other terms in the same lexical field or semantic network of related terms. The need for such a workbench derives from the fact that although current terminology management systems incorporate principles for organising the conceptual structure of terminologies, little or no concern is given to the formal representation of the conceptual systems behind the terminologies of the corresponding domains (Vouros and Eumeridou 2002). Moreover, from our experience, it appears that many tools needed to support our terminological work are either not present in existing commercialised terminology management systems or are very difficult to handle due to the particular organisation of the software workbenches.

This document provides a survey of all knowledge resources used for the development of the ontology and the multilingual terminological database in the FF POIROT project. Section 1 lists all fraudulent websites that have been used for modelling fraud schemes (section 1.1) as well as websites set up in the fight against fraud (section 1.2). Section 2 surveys the multilingual text corpus consisting of background reading texts (sections 2.1 and 2.2) as well as legislative texts (sections 2.3 and 2.4). In section 3 we summarise all dictionaries and glossaries that have been consulted by terminographers to further specify the content of the multilingual terminology base. Section 4 provides a survey of all ontologies according to which the FF POIROT application ontology will be structured. Finally, section 5 lists all stories used for knowledge analysis.

This deliverable reports on the difficulties encountered in WP4 with respect to the development of the FF POIROT multilingual terminological database. A first draft of this database was released in the month of March 2004. The second and final release appeared at the end of September 2004. This document will be structured as follows: in the first section, we reflect on the working method applied in WP4. We outline the encountered problems and propose a new methodology (called Termontograhy) which we presented at several conferences and workshops. The second section, explains the difficulties with respect to the compilation of the multilingual textual corpus used in WP4. The third section covers problems encountered with tools used for extracting and managing multilingual terminology. Finally, the fourth section lists the typical terminology problems encountered.

We start from a brief discussion of three shifts which have affected the discipline of terminology over the last decade: the shift towards computational terminology management, the linguistic shift in the theory of terminology and more recently the ontology shift. We briefly introduce the method of termontography, one of the results of the collaboration between terminologists from CVC Brussels and ontology engineers within the framework of the European project FF POIROT (IST 2001-38248). The discipline of terminology is widening its scope towards knowledge representation and knowledge management. On the one hand reflection is needed on how the discipline of terminology can contribute to new challenges like the(multilingual) semantic web. On the other hand we need to ask how terminology management can benefit from new developments in IT and NL processing, and more specifically how terminology management can use ontologies.

In this report we will discuss how the Centre for Terminology and Communication (CVC) may contribute to the development of an ontology related to financial fraud.

This deliverable discusses several possible techniques to align the FF POIROT multilingual terminological database developed by CVC with the lexon base developed by STAR Lab VUB. The aim of the alignment is to provide the lexon base with multilingual information. The database has been compiled with Trados MultiTerm 5.5. We will focus on the main export formats and explore the necessary modifications to be able to convert the terminology base to an ontology upload format set up by Star Lab VUB.

This document summarises our progress in WP4 with respect to the development of the multilingual terminological database (section 1). The second section deals with our experience regarding the Trados terminology workbench, the LinkFactory tool of Language and Computing and the terminology aligner of RACAI. Finally, in the third section we briefly discuss the problems that we encountered in setting up the FFPOIROT terminological database. A more elaborate discussion of these problems or difficulties will be reported in deliverable 4.2, which is to be expected at the end of month 24.

We intend to reflect on the problems encountered and methods applied when trying to represent multilingual and culture-specific knowledge in a terminological database which is set up to provide multilingual input for ontology engineers. The database consists of terminological information referring to units of understanding in the domain of value added tax. This database could be one of the main resources for the multilingual refinement within the DOGMA ontology engineering framework. We focus on how the method, called termontography, adds to the representation of multilingual and culture-specific knowledge in a European VAT regulatory ontology. What we propose is a unit of understanding approach which starts from a language-independent domain-specific framework arrived at with the help of domain specialists.

This document proposes a common workflow for the development of multilingual ontologies.

This document discusses our proposal for a termontography approach, an alternative approach in terminography which is being developed within the framework of FFPOIROT (section 1). The second section in this report will focus on the requirements concerning a software platform which is to support the termontography approach.

In February 2006 the first international conference on Multilingualism and Applied Comparative Linguistics (MACL) was held in Brussels, Belgium. The aim of the MACL conference was to bring together scholars from various branches of applied linguistics with a shared interest in cross-linguistic and cross-cultural communication. The conference thus fostered an exchange of knowledge and expertise among researchers from various disciplines, including educational linguistics, cultural linguistics, terminography, translation studies and studies of specialised languages. The present book is the second of two volumes containing a selection from the approximately 120 papers that were presented at that three-day event. The book comprises five chapters, reflecting different research perspectives on cross-linguistic and cross-cultural communication. The first chapter covers research articles on metaphors and planned languages. The second chapter comprises articles dealing with language attitudes, language proficiency and language practices in cross-linguistic and cross-cultural, communicative contexts. Chapter three features articles in the field of discourse-analysis research. In the fourth chapter research is presented that pertains to terminology and specialised languages. Finally, chapter five deals with translation studies.