Text-to-Stage: Spatial Layouts from Long-form Narratives
This work addresses the automation of spatial reasoning for media applications, but it is incremental as it builds on existing methods with specific enhancements.
The paper tackled the problem of inferring stage-play layouts from long-form narratives lacking explicit spatial cues, and introduced a training recipe that improved performance over vanilla models across multiple metrics, including character attribution and spatial plausibility.
In this work, we probe the ability of a language model to demonstrate spatial reasoning from unstructured text, mimicking human capabilities and automating a process that benefits many downstream media applications. Concretely, we study the narrative-to-play task: inferring stage-play layouts (scenes, speaker positions, movements, and room types) from text that lacks explicit spatial, positional, or relational cues. We then introduce a dramaturgy-inspired deterministic evaluation suite and, finally, a training and inference recipe that combines rejection SFT using Best-of-N sampling with RL from verifiable rewards via GRPO. Experiments on a text-only corpus of classical English literature demonstrate improvements over vanilla models across multiple metrics (character attribution, spatial plausibility, and movement economy), as well as alignment with an LLM-as-a-judge and subjective human preferences.