PhD student

Université Joseph Fourier
Laboratoire d’Informatique de Grenoble équipe SIGMA (LIG – SIGMA)
BP 72, 38402 Saint Martin d’Hères Cedex – France

Tel +33  4 76 63 55 68
Mobile +33 6 95 79 09 01
Minh-Tien.nguyen@imag.fr

Semantic Detection, Management and Analysis in scientific literature.

Editing process make an extensive use of meta-data to describe texts (XML/SGML). For now semantic structures in text content of scientific publishers are mostly related to structural information. With these structural markup  (e.g. a document, a title or a paragraph) in place, it is possible to handle article and chapter content in an automated way during a production and online publication process within a publishers business process management IT system.
But the meaning of content within these structures is still unknown and not available. Automatic detection and markup of information related to the semantic carried by a portion of text would be of a great use. Such information will leverage, both, the development of semantic data related products (e.g. exposing semantically enriched texts to readers) and also checking, validation and quality process (e.g assesing topics and meaning).

Concepts and ideas as well as the implementation of tools to achieve semantic detection, management and analysis are at the center of the PHD thesis. Several technics from the field of text miming and machine learning will be tested in this context. Deep language analysis techniques (entity recognition and semantic extraction) will also be explored to identify very particular feature related to scientific texts.

This work is founded by Springer and thus will be done in very close collaboration and tied to operative implementation of considered techniques. Experimentation could also be done on real material.

Keywords: quantitative and statistic approaches, machine learning statistics, entity recognition, natural language processing.Voici votre premier post ou article. Par défaut vos derniers articles viennent s’afficher l’un à la suite de l’autre à l’ouverture du site. Vous pouvez faire le choix d’une « static page », la page d’accueil par exemple en allant dans : settings / reading. Pour paramétrer votre site, veuillez éditer cette page, vous aurez accès à un espace d’administration.
Ce modèle est une adaptation du thème zeeMagazine pour le LIG. Vous avez à votre disposition d’autres thèmes.
Vous trouverez un exemple de site et un mode d’emploi (menu du haut) sur le site : PagePerso.

Pour toute demande ou remarque vous pouvez contacter : contact-si.lig@imag.fr