Rethinking Traditional Planning Assumptions to Facilitate Narrative Generation
Stephen G. Ware, R. Michael Young · National Conference on Artificial Intelligence · 2010
STRIPS-style planning has proven to be a helpful methodology for narrative generation, but certain assumptions about the process remain in use which inhibit the creation of interesting stories. The sequence of actions is more important than the initial and goal state of the world, so a narrative planner should first build a plot and then adapt the world to that plot. This is possible by relaxing the closed world assumption to allow revision to the initial and goal states. Adapting Planning for Story Writing The mature AI formalism of STRIPS-style planning has proven to be a useful tool for narrative generation. Research groups, especially those headed by Young, Cavazza, Mateas, and Riedl have produced a number of story writing systems based on planners. The STRIPS planning representation is a natural one for story writing because it resembles existing narratology models (Cavazza and Pizzi 2006) and can be easily mapped onto psychological models of story comprehension (Young 1999). Classical planning was first developed for solving real world problems in which the initial and goal states were known and a sequence of actions to achieve the goal was desired. Most classical domains are assumed to be fully observable, deterministic, static, and discrete (Russell and Norvig 2003). These assumptions are unrealistic for the real world. However, these assumptions can be reasonable for story generation. The author is usually free to take a God’s eye view of the story world (fully observable), determine the exact outcomes of actions (deterministic), control the environment (static), and estimate real world concepts to an arbitrary degree of accuracy (discrete). Ironically, classical planning is much better suited to story generation than to its original purpose because of these assumptions. But some traditional assumptions about input and output no longer hold, and those should be critically examined.