FACTTRACK: Time-Aware World State Tracking in Story Outlines

We propose FactTrack, an algorithm that decomposes complex narratives into atomic facts and models dynamic world states for contradiction detection. This work addresses the challenging problem of inner-plot factual consistency checking and temporal fact tracking in narrative structures like story outlines.

The key innovation is our temporal fact tracking mechanism that maintains a dynamic world state throughout the narrative timeline, enabling detection of contradictions that emerge from the evolving story context.

Download paper here

Recommended citation:

@inproceedings{lyu2025facttrack,
  title={FACTTRACK: Time-Aware World State Tracking in Story Outlines},
  author={Lyu, Zhiheng and Yang, Kevin and Kong, Lingpeng and Klein, Dan},
  booktitle={Proceedings of the 2025 Conference of the North American Chapter of the Association for Computational Linguistics},
  year={2025}
}