Information regarding the Mimesis system and related projects can be found on the Liquid Narrative Group website.
Mimesis makes use of planning algorithms to calculate actions and events in a story. A planner identifies an initial plan starting from a set of initial conditions and reaching a final situation after a number of steps defined by the author (i.e. the protagonist reaches a certain goal, while another character is in a certain state).
If the user's intervention threatens the precalculated plan, the system performs one of the two alternative strategies:
- Accommodation: the planner recalculates another plan that integrates the user's action.
- Intervention: the planner changes the user's action by making it fail. For example, a coin machine will not work (so that the coin is not lost) or a shooting fails (so that the targeted character remains alive).
Mimesis in integrated within a Real-Time 3D environment, via the Unreal Tournament game engine.
Result Description (end user perspective)
Limited information regarding authoring is provided. A more recent paper however is concerned with the specific issue of authoring. It describes a plan-authoring Graphical User Interface called Bowman. This authoring tool allows to (or will allow to):
- Visualize and navigate the entire plan space of a story
- Provide a meta-level control over the generated plans, in order to improve its quality
- Enrich the planning domain according to informations provided by the planner (i.e. the system could suggest when user's alternative actions should be added).
Mimesis is based on a solild theoretical basis concerning planning. From a core engine for narrative planning, several research directions have been undertaken, regarding suspense modeling, mixed initiative planning, camera control, planning web services, etc.
As for several planning-based interactive narrative engines, the appraoch is focused on the final state of the plan, less on the plan quality itself. Recent research on suspense modeling tend to ovecome the limitation. No demo is available; The system does not seem to have been used by authors in practical creations.
- Bae, B.-C. & Young, R. M. (2008). A Use of Flashback and Foreshadowing for Surprise Arousal in Narrative Using a Plan-Based Approach. ]. In Spierling & Szilas (Eds.) Proc. Interactive Storytelling - ICIDS 2008 (pp 156-167). Springer Verlag.
- Cheong, Y-G. & Young, R. M. (2008). Narrative Generation for Suspense: Modeling and Evaluation. In Spierling & Szilas (Eds.) Proc. Interactive Storytelling - ICIDS 2008 (pp 144-155). Springer Verlag.
- Jhala, A. and Young, R. M. (2005). A Discourse Planning Approach for Cinematic Camera Control for Narratives in Virtual Environments, in Proceedings of the National Conference of the American Association for Artificial Intelligence.
- Riedl, M., Saretto, C.J. and Young, R. M. (2003). Managing interaction between users and agents in a multiagent storytelling environment, in the Proceedings of the Second International Conference on Autonomous Agents and Multi-Agent Systems.
Supporting Narrative Theories
The computational models of suspense associated to Mimesis is based on the concept of a model of the reader, which consists in evaluating the inferences performed by the user during the experience (see the textual cooperation). More precisely, the model is derived from narrative theories of suspense.
Mimesis makes use of planning tools and techniques.