Agents' Room: Narrative Generation through Multi-step Collaboration

2024-10-03

arXiv.org

AI Summary

Existing large language models (LLMs) suffer from low narrative quality, heavy reliance on complex prompts, and poor controllability and interpretability in long-form storytelling. To address these limitations, we propose a narrative-theory-inspired multi-agent collaborative framework that decomposes story generation into specialized subtasks—plot development, character modeling, and linguistic realization—each handled by dedicated, cooperating agents. Our key contributions are: (1) the first narrative-driven multi-agent division-of-labor paradigm; (2) “Tell Me A Story”, the first high-quality prompt–story paired dataset; and (3) a hybrid evaluation framework for long narratives, integrating expert human assessment with automated metrics. Experiments demonstrate statistically significant improvements (p < 0.01) over strong baselines across coherence, creativity, and character consistency. Moreover, our approach enables controllable, interpretable, and iterative narrative generation.

Technology Category

• Humans and AI: Game Design — Procedural Content Generation & Storytelling
• Natural Language Processing: Generation
• Multiagent Systems: Teamwork

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems

Abstract

Writing compelling fiction is a multifaceted process combining elements such as crafting a plot, developing interesting characters, and using evocative language. While large language models (LLMs) show promise for story writing, they currently rely heavily on intricate prompting, which limits their use. We propose Agents' Room, a generation framework inspired by narrative theory, that decomposes narrative writing into subtasks tackled by specialized agents. To illustrate our method, we introduce Tell Me A Story, a high-quality dataset of complex writing prompts and human-written stories, and a novel evaluation framework designed specifically for assessing long narratives. We show that Agents' Room generates stories that are preferred by expert evaluators over those produced by baseline systems by leveraging collaboration and specialization to decompose the complex story writing task into tractable components. We provide extensive analysis with automated and human-based metrics of the generated output.

Problem

• Decomposes narrative writing into specialized subtasks
• Addresses limitations of LLMs in story generation
• Improves story quality through multi-agent collaboration

Innovation

• Decomposes narrative writing into specialized subtasks
• Uses collaborative agents for complex story generation
• Introduces novel evaluation framework for long narratives