A Computational Model of Narrative Generation for Suspense

Abstract

The generation of stories by computers, with applications ranging from computer games to education and training, has been the focus of research by computational linguists and AI researchers since the early 1970s. Although several approaches have shown promise in their ability to generate narrative, there has been little research on the generation of stories that evoke specific cognitive and affective responses in their readers. The goal of this research is to develop a system that produces a narrative designed specifically to evoke a targeted degree of suspense, a significant contributor to the level of engagement experienced by users of interactive narrative systems. The system that I present takes as input a plan data structure representing the goals of a storyworld's characters and the actions they perform in pursuit of them. Adapting theories developed by cognitive psychologists, my system uses a plan-based model of narrative comprehension to determine the final content of the story in order to manipulate a reader's level of suspense in specific ways. In this thesis, I outline the various components of the system and describe an empirical evaluation that I used to determine the efficacy of my techniques. The evaluation provides strong support for the claim that the system is effective in generating suspenseful stories.