Interoperable annotation of (co)references in the Democrat project

摘要

This paper proposes XML-TEI-URS, a generic TEI-based format for the annotation of coreferences in arbitrary corpora. This proposal is made in the context of Democrat, a French Agence Nationale de la Recherche project that aims to produce a large corpus of written French with coreference annotations, in an attempt to design a corpus that is usable both by humans and automated tools and as compatible as possible with future concurrent annotations.