Nathalie Vallès-Parlangeau

Personalized information access through flexible and interoperable profiles

When searching information, any user has to face huge cognitive efforts to obtain accurate and relevant results. The search task includes a set of complementary sub-tasks in which the user needs to be necessarily involved. But, the real place of the users is not obvious without an effective knowledge of their context, environment, and so on. So we assume that a better knowledge of the user and of available information should make it possible to implement techniques aimed at adapting the retrieved information contents, as well as the search process itself. This personalization mainly relies on the definition of profiles. Since applications principally manage specific user/information profiles (structure and content), we propose in this paper a generic and a flexible profile model. This latter facilitates the construction and the interoperability of various profiles coming from different applications and/or having different structure/content. This paper presents the way the different resources (user, information...) can be modeled within the information search process and its related tasks. Then, we discuss the usefulness of profiles in such processes/tasks. Finally we present the generic and the flexible profile model we propose.

Formal modeling of multistructured documents

The quantity of digital documents available is still growing. The various contexts of use of such documents need several kinds of descriptions of their contents and structures. Thus a same document can be described according to several concurrent structures. Designing models and tools to exploit these various kinds of structures simultaneously presents a real challenge. In this way we have built document repositories to achieve this aim. Indeed, we proposed fragmentation techniques to manage the various issues related to the management of multistructured documents (representation, storage, reconstruction, and management of concurrent structures). This paper is dedicated to the presentation of the formal model. We propose to describe with precision and concision the various concepts related to the multi-structured documents as well as rules related to the organization of these documents.

Modeling and Exploitation of Multistructured Documents

More than half of information used in organizations is stored in documents. The diversity of their uses implies several descriptions and thus several decomposition needs of these documents (and so several structures). Hence, designing models and tools to exploit simultaneously these various kinds of structures presents a real challenge. To fulfill these needs, we propose a modeling approach based on fragmentation techniques to manage the various issues related to the management of multistructured documents (representation, storage, reconstruction, and management of concurrent structures). Three exploitation modes are proposed according to the different kinds of information searched: information, data interrogation and multidimensional analysis.

Management of document multistructurality: Case of document versions

Documents, either complex or not, can have several structures. This multistructurality can be due to different kinds of structures as well as different versions of a same document. Moreover, the document contents are dynamic and they change in the course of time. In this paper we show how the different document versions can be handled according to a model of multistructured documents. This model should ensure the representation of each version, the management of common contents and the monitoring of version evolution. Thus, our design methodology is based on a fragmentation according to which the document is fragmented in structuring nodes (elements, attributes, metadata) while keeping the relations between these nodes. These relations ensure the rebuilding of each document and the management of document versions.

Meta-mining Evaluation Framework: A Large Scale Proof of Concept on Meta-learning

This paper aims to provide a unified framework for the evaluation and comparison of the many emergent meta-mining techniques. This framework is illustrated on the case study of the meta-learning problem in a large scale experiment. The results of this experiment are then explored through hypothesis testing in order to provide insight regarding the performance of the different meta-learning schemes, advertising the potential of our approach regarding meta-level knowledge discovery.

Usagers et recherche d'information: Modélisation des usagers pour la recherche d'information adaptative

User profiles are the main component of adaptive information retrieval systems. We are identifying in this paper main limits of the usage of user profiles. We are discussing about these limits in underlying some solutions and taking into account two points of view: the application point of view and the user point of view.

Accès au contenu des thèses numériques par leur structure sémantique

Les projets de bibliotheques numeriques actuels offrent a l’utilisateur l’acces aux theses a partir d’une recherche qui ne permet pas d’extraire les parties pertinentes de la these et ne renvoie que la these integrale. Ainsi, l’utilisateur doit lire des chapitres entiers pour connaitre les parties qui correspondent a son besoin. Le projet CITHER (Consultation en texte Integral des THeses En Reseau) de l’INSA de Lyon dans lequel s’inscrit cette etude, porte sur la mise en ligne des theses. Nous proposons de permettre un acces pertinent au contenu des theses grâce a l’utilisation de « tags semantiques » rajoutes, par le doctorant, au sein de sa these lors de la redaction. L’exploitation de ces tags permet de cibler la recherche et ainsi mieux satisfaire l’utilisateur. Notre travail porte d’une part sur la constitution d’une base de concepts utilises pour le « tagage » de la these et, d’autre part, sur la definition d’un nouveau modele de documents a partir des differentes structures de la these.

Une approche multivue pour la gestion des documents multistructurés

La gestion flexible des documents numeriques se situe aujourd’hui au centre des preoccupations des decideurs (production collaborative, partage, personnalisation, actualisation, agregation, etc.). Il est evident ainsi qu’un meme document peut avoir plusieurs descriptions et donc plusieurs decompositions et plusieurs structures selon plusieurs contextes. Il apparait que la gestion de cette multistructuralite par des techniques de classification peut etre un element cle pour garantir un stockage efficace et une exploitation flexible des contenus des documents multimedias. L’objectif de cet article est de presenter notre approche pour la gestion de la multistructuralite dans le cadre d’un entrepot de documents multimedia. Au travers des methodologies de modelisation et d’exploitation flexible, nous permettons une prise en compte des differentes structures rattachees a un meme document.