A Distributed Architecture for Interactive Parse Annotation

In this paper we describe a modular system architecture for distributed parse annotation using interactive correction. This involves interactively adding constraints to an existing parse until the returned parse is correct. Using a mixed initiative approach, human annotators interact live with distributed ccg parser servers through an annotation gui. The examples presented to each annotator are selected by an active learning framework to maximise the value of the annotated corpus for machine learners. We report on an initial implementation based on a distributed workflow architecture.