Papers by Mounira Harzallah
Proceedings of the 10th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management, 2018
This paper aims to use term clustering to build a modular ontology according to core ontology fro... more This paper aims to use term clustering to build a modular ontology according to core ontology from domainspecific text. The acquisition of semantic knowledge focuses on noun phrase appearing with the same syntactic roles in relation to a verb or its preposition combination in a sentence. The construction of this cooccurrence matrix from context helps to build feature space of noun phrases, which is then transformed to several encoding representations including feature selection and dimensionality reduction. In addition, the content has also been presented with the construction of word vectors. These representations are clustered respectively with K-Means and Affinity Propagation (AP) methods, which differentiate into the term clustering frameworks. Due to the randomness of K-Means, iteration efforts are adopted to find the optimal parameter. The frameworks are evaluated extensively where AP shows dominant effectiveness for cooccurred terms and NMF encoding technique is salient by its promising facilities in feature compression.
This paper aims to analyze and adopt the term clustering method for building a modular ontology a... more This paper aims to analyze and adopt the term clustering method for building a modular ontology according to its core ontology. The acquisition of semantic knowledge focuses on noun phrase appearing with the same syntactic roles in relation to a verb or its preposition combination in a sentence. The construction of this co-occurrence matrix from context helps to build feature space of noun phrases, which is then transformed to several encoding representations including feature selection and dimensionality reduction. In addition, word embedding techniques are also presented as feature representation. These representations are clustered respectively with K-Means, K-Medoids, Affinity Propagation, DBscan and co-clustering algorithms. The feature representation and clustering methods constitute the major sections of term clustering frameworks. Due to the randomness of clustering approaches, iteration efforts are adopted to find the optimal parameter and provide convinced value for evalua...
Dans le contexte du web semantique et des masses de donnees, les ontologies necessaires a faire c... more Dans le contexte du web semantique et des masses de donnees, les ontologies necessaires a faire communiquer les objets du web seront probablement de grande taille et construites a partir des ressources diverses et heterogenes. Leur validation sera plus que jamais primordiale. Dans cet article, nous mettons en exergue les liens entre deux types problemes pouvant nuire a la qualite d’une ontologie : l’insatisfiabilite d’artefacts et le probleme de contradiction sociale. Nous proposons des anti-patrons partiels et une heuristique pour aider a identifier les problemes de contradiction sociale, tout en minimisant l’intervention humaine.
Partant du postulat que l’etude de l’interaction humain-humain peut servir a modeliser l’interact... more Partant du postulat que l’etude de l’interaction humain-humain peut servir a modeliser l’interaction humain-machine, nous etudions un corpus de dialogues de recherche collaborative de documents. De cette analyse est ne un modele de la tâche dans le cadre d’une interaction humain-humain. L’utilite de ce modele est discute pour le cas d’une interaction humain-machine.
Proceedings of the 10th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management, 2018
Hypernym relation extraction is considered the backbone of building ontologies. Hearst patterns a... more Hypernym relation extraction is considered the backbone of building ontologies. Hearst patterns are the most popular patterns used to extract hypernym relation. They include POS tags and lexical information, and they are applied on a shallow parsed corpora. In this paper, we propose a new formalization of Hearst patterns using dependency parser, called Dependency Hearst patterns. This formalization allows them to match better complex or ambiguous sentences. To evaluate our proposal, we have compared the performance of Dependency Hearst patterns to that of the lexico-syntactic Hearst patterns, applied on a music corpus. Dependency Hearst patterns yield a better result than lexico-syntactic patterns for extracting hypernym relations from the corpus.
Proceedings of The 12th International Workshop on Semantic Evaluation, 2018
In this paper, we present our proposed system (EXPR) to participate in the hypernym discovery tas... more In this paper, we present our proposed system (EXPR) to participate in the hypernym discovery task of SemEval 2018. The task addresses the challenge of discovering hypernym relations from a text corpus. Our proposal is a combined approach of path-based technique and distributional technique. We use dependency parser on a corpus to extract candidate hypernyms and represent their dependency paths as a feature vector. The feature vector is concatenated with a feature vector obtained using Wikipedia pre-trained term embedding model. The concatenated feature vector fits a supervised machine learning method to learn a classifier model. This model is able to classify new candidate hypernyms as hypernym or not. Our system performs well to discover new hypernyms not defined in gold hypernyms.
Un Outil De Gestion Des Competences Acquises et Requises Implante a Tremery Dans Le Groupe Psa
La compétence et la connaissance sont deux concepts qui nous semblent fortement conjoints, cepend... more La compétence et la connaissance sont deux concepts qui nous semblent fortement conjoints, cependant, ils sont rarement étudiés et gérés ensemble. Nous cherchons donc à identifier les liens et frontières qui peuvent exister entre eux. Ceci a pour objectif de développer un modèle de représentation et de gestion, intégré aux connaissances et aux compétences. Dans cet article, est tout d'abord présentée, une synthèse sur les concepts de compétence et de connaissance. Ensuite, les modèles et outils de gestion de ces concepts sont exposés. Puis, le modèle CKIM (Competency and Knowledge Integrated Model) développé, est défini. Les utilités de ce modèle et son exploitation sont discutées en quatrième partie. La dernière partie représente un prototype d'implantation du modèle CKIM réalisé sur le serveur de connaissances ATHANOR.
Nous nous intéressons au raisonnement sur les compétences des ressources humaines pour simplifier... more Nous nous intéressons au raisonnement sur les compétences des ressources humaines pour simplifier leur gestion. Dans cet article, nous proposons une méthode de raisonnement pour l'aide à l'identification des compétences d'un individu. Un processus de knowledge-mining défini par analogie avec l'extraction de règles d'association en data-mining est proposé afin d'induire une base de règles à partir d'une base de connaissances sur le domaine. De plus, un prototype a été développé pour expérimenter notre approche sur un exemple académique.
Procedia Computer Science, 2015
Comparing objects described (or annotated) with an ontology is quite important in several applica... more Comparing objects described (or annotated) with an ontology is quite important in several application domains. In our previous work we have shown that when objects are annotated with single concepts or sets of concepts, most of the semantic measures found in literature can be rewritten by using a unified form and the notion of approximated information content. In this paper, we argue how this unified form and the notion of approximated information content can be extended to semantic graphs, being the latter resulting from objects annotated with graphs of concepts connected via relationships found in the ontology. We show that open issues and distortions found in the relevant state of the art are put under control. The resulting unified framework (covering all types of annotations, from single concepts to graphs of concepts) is therefore relevant for practicing, in a unique environment, several semantic measures, as required for building a new one or selecting a measure for given application objectives.
Building ontologies and debugging them is a timeconsuming task. Over the recent years, several ap... more Building ontologies and debugging them is a timeconsuming task. Over the recent years, several approaches and tools for the automatic construction of ontologies from textual resources have been proposed. But, due to the limitations highlighted by experimentations in real-life applications, different researches focused on the identification and classification of the errors that affect the ontology quality. However, these classifications are incomplete and the error description is not yet standardized. In this paper we introduce a new framework providing standardized definitions which leads to a new error classification that removes ambiguities of the previous ones. Then, we focus on the quality of automatically built ontologies and we present experimental results of our analysis on an ontology automatically built by Text2Onto for the domain of composite materials manufacturing.
Enterprise Interoperability II
While current approaches to ontology mapping produce good results by mainly relying on label and ... more While current approaches to ontology mapping produce good results by mainly relying on label and structure based similarity measures, there are several cases in which they fail to discover important mappings. In this paper we describe a novel approach to ontology mapping, which is able to avoid this limitation by using background knowledge. Existing approaches relying on background knowledge typically have one or both of two key limitations: 1) they rely on a manually selected reference ontology; 2) they suffer from the noise introduced by the use of semi-structured sources, such as text corpora. Our technique circumvents these limitations by exploiting the increasing amount of semantic resources available online. As a result, there is no need either for a manually selected reference ontology (the relevant ontologies are dynamically selected from an online ontology repository), or for transforming background knowledge in an ontological form. The promising results from experiments on two real life thesauri indicate both that our approach has a high precision and also that it can find mappings, which are typically missed by existing approaches.
Digital Resource Discovery: Semantic Annotation and Matching Techniques
Enterprise Interoperability, 2007
Proceedings of the International Conference on Knowledge Engineering and Ontology Development, 2014
Ontologies play a major role in the development of personalized and interoperable applications. H... more Ontologies play a major role in the development of personalized and interoperable applications. However, validation of ontologies remains a critical open issue. Validation is fundamentally driven by an "ontology evaluation", often referred to as "quality evaluation", as better explained in the Introduction. This paper reports an experience designed on our previous work on quality evaluation and using ontologies automatically generated from some textual resources. In the previous work, we have proposed a standard typology of problems impacting (negatively) on the quality of one ontology (named quality problems). The experience shows how our previous work can be practically deployed. One a posteriori analysis of experience results and lessons learnt presented in the paper make explicit and concrete key contributions to validation. Finally, conclusions highlight both limitations of the experience and research perspectives.
Revue nationale des …, 2006
Résumé. Les ontologies sont au coeur du processus de gestion des connaissances. Différentes mesur... more Résumé. Les ontologies sont au coeur du processus de gestion des connaissances. Différentes mesures sémantiques ont été proposées dans la littérature pour évaluer quantitativement l'importance de la liaison sémantique entre paires de concepts. Cet ...
EMOI-INTEROP'05, …, 2005
Abstract. Ontologies are in the heart of the knowledge management process. Different semantic mea... more Abstract. Ontologies are in the heart of the knowledge management process. Different semantic measures have been proposed in the literature to evaluate the strength of the semantic link between two concepts or two groups of concepts from either two different ...
Computers in Industry, 2010
The Unified Enterprise Modelling Language (UEML) aims at supporting integrated use of enterprise ... more The Unified Enterprise Modelling Language (UEML) aims at supporting integrated use of enterprise and IS models expressed using different languages. To achieve this aim, UEML offers a hub through which modelling languages can be connected, thereby paving the way for also connecting the models expressed in those languages. This paper motivates and presents the most central parts of the UEML approach: a structured path to describing enterprise and IS modelling constructs; a common ontology to interrelate construct descriptions at the semantic level; a correspondence analysis approach to estimate semantic construct similarity; a quality framework to aid selection of languages; a meta-meta model to integrate the different parts of the approach; and a set of tools to aid its use and evolution. The paper also discusses the benefits of UEML and points to paths for further work.
ACTES DE LA CONFÉRENCE
De l'Ingénierie des Connaissances à la Gestion des Compétences Giuseppe Berio1, Mounira Harz... more De l'Ingénierie des Connaissances à la Gestion des Compétences Giuseppe Berio1, Mounira Harzallah² 1 Dipartimento di Informatica, Università di Torino ... Garro & Palopoli (2003) considèrent un système de «e-learning» qui peut de l'ingénierie des connaissances à la gestion ...
Knowledge management for competence management
Journal of Universal Knowledge Management, 2005
L’ingenierie des connaissances (extraction, modelisation, capitalisation, exploitation...) a conn... more L’ingenierie des connaissances (extraction, modelisation, capitalisation, exploitation...) a connu plusieurs mutations en s’adaptant au cours du temps a l’evolution des connaissances. Elle a notamment du prendre en compte une evolution dans le temps des ressources des connaissances (experts, livres, bases de donnees, reseaux sociaux, tweeters, web des donnees...), de leurs formes (implicite, explicite, structuree, semi ou non structuree), de leurs utilisateurs (organisation, apprenant, utilisateur du web...), des supports de leur utilisation (livres, bases de donnees, systemes a bases de connaissances, applications du web semantique...), du volume et de la vitesse de multiplication de leurs ressources, des techniques de leur extraction, des langages de leur representation... Dans ces differentes evolutions, l’ontologie a ete reconnue comme une representation semantique pertinente des connaissances. Je me suis interessee depuis plus de 13 ans, aux problematiques liees aux ontologies,...
Uploads
Papers by Mounira Harzallah