Reproduce “STORYTELLER: An Enhanced Plot-Planning Framework for Coherent and Cohesive Story Generation” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Jiaming Li, Yukun Chen, Ziqiang Liu, Minghuan Tan, Lei Zhang, Yunshui Li, Run Luo, Longze Chen and 5 more
STORYTELLER, a framework from Shenzhen Institutes of Advanced Technology and the University of Chinese Academy of Sciences, enhances automatic story generation by employing a top-down plot-planning strategy with dynamic interaction between plot nodes and a narrative knowledge graph. It significantly improved narrative coherence, creativity, and engagement, achieving an 89.4 overall quality score and up to a 91% human preference win rate against GPT-4o.
Key insights
- Integrating a dynamic plot-planning mechanism with granular control through Subject-Verb-Object (SVO) plot nodes helps maintain structural integrity and thematic continuity.
- Continuously interacting with a Narrative Entity Knowledge Graph (NEKG) allows for robust tracking of character interactions, object dynamics, and logical progression, aiding in plot generation and reasoning.
- An iterative review process for plot nodes, informed by past events and entity relationships, is crucial for enriching narrative content, enhancing creativity, and ensuring coherence across chapters.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.