talnarchives

Une archive numérique francophone des articles de recherche en Traitement Automatique de la Langue.

SIDRES : A Novel Annotation Tool For The Automatic Detection of Semantic Entities

Julieta Murata, Rémy Carrette, Pierre Jourlin

Abstract : We present a novel annotation tool called SIDRES (Système Interactif de Détection et de Reconnaissance d’Entités Sémantiques [Interactive System for the Detection and Identification of Semantic Entities]). SIDRES provides an annotation environment for classifying text units through ad hoc categories. These categories can be coupled with contexts, so as to provide a means for the disambiguation of formally identical units assigned to different categories. SIDRES was developed as part of an industrial partnership between the LIA (Laboratoire d’Informatique d’Avignon [Research Institute of Informatics at the University of Avignon]) and a French company in the eHealth sector. This partnership was created within the framework of a technology-transfer project promoted by the SATT Sud-Est, whose core mission is bringing together industry and research institutions.

Keywords : medical reports, annotation, locally ambiguous grammar