Winnow: A Domain-Specific Language for Incremental Story Sifting

Authors

  • Max Kreminski University of California, Santa Cruz
  • Melanie Dickinson University of California, Santa Cruz
  • Michael Mateas University of California, Santa Cruz

Keywords:

Story Sifting, Emergent Narrative, Interactive Narrative, Logic Programming, Domain-specific Languages, Datalog

Abstract

Story sifters attempt to automatically or semi-automatically extract nuggets of compelling narrative content from vast chronicles of game or simulation events. Though sifting has successfully been used to enable novel computational narrative play experiences, its utility is limited by the fundamentally retrospective nature of existing sifters, which can only recognize storyful event sequences once they have fully played out. To address this limitation, we introduce Winnow: a domain-specific language for specifying story sifting patterns that can be executed incrementally to detect potentially storyful event sequences while they are still playing out. We evaluate Winnow by applying it to several specific use cases and show that it is well-suited to the implementation of prospective as well as retrospective narrative intelligence.

Downloads

Published

2021-10-04

How to Cite

Kreminski, M., Dickinson, M., & Mateas, M. (2021). Winnow: A Domain-Specific Language for Incremental Story Sifting. Proceedings of the AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment, 17(1), 156-163. Retrieved from https://ojs.aaai.org/index.php/AIIDE/article/view/18903