U-Director

From IDSwiki
Jump to: navigation, search

IRIS Wiki - IS Systems - U-Director

Availability

Non distributed prototype.

Technical Description

Utility-based Director Agent (U-Director) is a director agent that monitors the storytelling according to narrative objectives, user states and storyworld states. It offers a narrative architecture that uses the dynamic decision network (Dean & Kanazawa, 1989). It selects the action to be performed during the unfolding of a story according to players' interaction and the storytelling. Narrative utility is maximized, according to various criteria such as plot progress, narrative flow, location flow, etc.

Result Description (end user perspective)

U-DIRECTOR had been implemented in the Crystal Island storytelling environment. It consists in unfolding the specific story on Crystal Island.

"The user’s character awakens from a good night’s sleep and the adventure begins. At this juncture, several elements in the plot graph are available for her to address. For example, she could find her father paralyzed in his bedroom, she might notice that a plate with leftover food is in his office, or she might discover other facts such as that her good friend, Teresa, has also been stricken with the mysterious illness."

Strong Points

U-DIRECTOR exploits recent advances in approximate Bayesian inference via stochastic sampling.

It is an emergent narrative approach that is directed by storytelling means.

Limitations

Results of the study indicate that the clustering algorithm’s running time has the greatest variability and may not satisfy the performance requirements of interactive narrative.

Main Publications

  • Mott, B. W. and Lester, J. C. (2006). U-director: a decision-theoretic narrative planning architecture for storytelling environments. In Proceedings of the Fifth international Joint Conference on Autonomous Agents and Multiagent Systems (Hakodate, Japan, May 08 - 12, 2006). AAMAS '06. ACM, New York, NY, 977-984. [1]
  • Dean, T. and Kanazawa, K. (1989). A model for reasoning about persistence and causation. Computational Intelligence, 5(3), 142-150. [2]

Supporting Narrative Theories

-

Computational Model

Bayesian networks.