Emotion-Driven Narrative Generation

Abstract

While a number of systems have been developed that can generate stories, the challenge of generating stories that elicit emotions from human audiences remains an open problem. With the development of models of emotion, it would be possible to use these models as means of evaluating stories for their emotional content. In this chapter, we discuss Dramatis, a model of suspense. This model measures the level of suspense in a story by attempting to determine the best method for the protagonist to avoid a negative outcome. We discuss the possibilities for Dramatis and other emotion models for improving intelligent generation of narratives.