Humans seem to be wired to understand the world and other people in terms of narrative representations: who did what? why did she do it?, and what is she likely to do next?. Thus when people share experiences in natural settings, they tell stories. In this talk we discuss our work on storytelling agents who retell existing stories, such as personal narratives posted on weblogs. We describe our method for representing the deep semantic structure of these stories, and how these representations allow us to produce thousands of different retellings of the same story, using narratologically-inspired verbal parameters and personality-based nonverbal parameters.

Marilyn Walker is a professor of  computer science and head of the Natural Language and Dialogue Systems Lab at the University of California, Santa Cruz. Her research focuses on computational models of dialogue, including work on dialogue in storytelling, argumentative dialogues, interactive games, and dialogue systems.  Before coming to Santa Cruz, Marilyn was a Professor of Computer Science at University of Sheffield. From 1996 to 2003, she was a Principal Member of Research Staff in the Speech and Information Processing Lab at AT&T Bell Labs and AT&T Labs Research. She has published more than 200 papers, and has 10 U.S. patents. She received her Ph.D. in computer science from University of Pennsylvania (1993). Her H-index, a measure of research impact, is 50.