The primary objective of this project is to develop new, cognitively informed computational models of the generation of narrative. Motivated by narrative theoretic models of narrative structure and psychological models of narrative comprehension, we are developing techniques for creating stories and the techniques needed to convey them to users. These techniques will use these models to search for narratives that are at once coherent and effective at communicating the underlying story structure. The research program has three major thrusts:

1. Integrating generative models of character plans with narrative theoretic structural models of story to create storylines that reflect both rich character goal structures and recognizable narrative elements.
2. Developing methods for shot and shot sequence selection that build on pragmatic models from linguistic communication to effectively convey characters’ plans and goals.
3. Developing and then evaluating a system that integrates these parts to search for narratives that are both coherent and effective.

We are providing a comprehensive account of the effectiveness of our models by conducting extensive empirical studies.

The project extends our previous work developing computational models of narrative, focusing on elements of the modeling and effective communication of character goals in narrative. We’;re exploring the notion that computational models of the mental processes performed by people when experiencing film or machinima can inform an automatic process used to generate the films themselves.

Team members