An Incremental Architecture for the Semantic Annotation of Dialogue Corpora with High-Level Structures. A case of study for the MEDIA corpus.

Lina Maria Rojas-Barahona and Matthieu Quignard


The semantic annotation of dialogue corpora permits building efficient language understanding applications for supporting enjoyable and effective human-machine interactions. Nevertheless, the annotation process could be costly, time-consuming and complicated, particularly the more expressive is the semantic formalism. In this work, we propose a bootstrapping architecture for the semantic annotation of dialogue corpora with rich structures, based on Dependency Syntax and Frame Semantics.