An intelligent authoring environment for abstract semantic representations of cultural object descriptions

Abstract

In this paper we describe an authoring environment for the creation of culturaldomain ontologies and the associated linguistic and profile annotations, for dynamically generating adaptable naturallanguage descriptions of the cultural objects in the ontology. Adaptation is achieved at the expense of considerable authoring effort, since it relies on providing numerical parameters for each ontological entity. To assist the authoring process, we provide an intelligent authoring back-end that completes manually authored models by inferring missing values. This intelligent authoring support facility, combined with immediate previews, can considerably reduce the effort required to create a fully functional model as the author can iterate through cycles of providing information, previewing the generated text, and only elaborating the model where the text is unsatisfactory.