S^2tory: Story Spine Distillation for Movie Script Summarization
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911646622941184 |
|---|---|
| author | Lu, Mingzhe Liu, Yanbing Wang, Qihao Zhang, Jiarui Wu, Jiayue Hu, Yue Li, Yunpeng Xu, Yangyan |
| author_facet | Lu, Mingzhe Liu, Yanbing Wang, Qihao Zhang, Jiarui Wu, Jiayue Hu, Yue Li, Yunpeng Xu, Yangyan |
| contents | Movie scripts pose a fundamental challenge for automatic summarization due to their non-linear, cross-cut narrative structure, which makes surface-level saliency methods ineffective at preserving core story progression. To address this, we introduce S^2tory (Story Spine Distillation), a narratology-grounded framework that leverages character development trajectories to identify plot nuclei, the essential events that drive the narrative forward, while filtering out peripheral satellite events that merely enrich atmosphere or emotion. Our Narrative Expert Agent (NEAgent) performs theory-constrained reasoning, whose distilled knowledge conditions a small model to identify plot nuclei. Another model then uses these plot nuclei to generate the summary. Experiments on the MovieSum dataset demonstrate state-of-the-art semantic fidelity at approximately 3.5x compression, and zero-shot evaluation on BookSum confirms strong out-of-domain generalization. Human evaluation further validates that narratological theory provides an indispensable foundation for modeling complex, non-linear narratives. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_03244 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | S^2tory: Story Spine Distillation for Movie Script Summarization Lu, Mingzhe Liu, Yanbing Wang, Qihao Zhang, Jiarui Wu, Jiayue Hu, Yue Li, Yunpeng Xu, Yangyan Computation and Language Artificial Intelligence Movie scripts pose a fundamental challenge for automatic summarization due to their non-linear, cross-cut narrative structure, which makes surface-level saliency methods ineffective at preserving core story progression. To address this, we introduce S^2tory (Story Spine Distillation), a narratology-grounded framework that leverages character development trajectories to identify plot nuclei, the essential events that drive the narrative forward, while filtering out peripheral satellite events that merely enrich atmosphere or emotion. Our Narrative Expert Agent (NEAgent) performs theory-constrained reasoning, whose distilled knowledge conditions a small model to identify plot nuclei. Another model then uses these plot nuclei to generate the summary. Experiments on the MovieSum dataset demonstrate state-of-the-art semantic fidelity at approximately 3.5x compression, and zero-shot evaluation on BookSum confirms strong out-of-domain generalization. Human evaluation further validates that narratological theory provides an indispensable foundation for modeling complex, non-linear narratives. |
| title | S^2tory: Story Spine Distillation for Movie Script Summarization |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2605.03244 |