[2507.15518] HAMLET: A Hierarchical and Adaptive Multi-Agent Framework for Live Embodied Theatrics

Computer Science > Artificial Intelligence arXiv:2507.15518 (cs) [Submitted on 21 Jul 2025 (v1), last revised 3 Mar 2026 (this version, v4)] Title:HAMLET: A Hierarchical and Adaptive Multi-Agent Framework for Live Embodied Theatrics Authors:Shufan Jiang, Sizhou Chen, Chi Zhang, Xiao-Lei Zhang, Xuelong Li View a PDF of the paper titled HAMLET: A Hierarchical and Adaptive Multi-Agent Framework for Live Embodied Theatrics, by Shufan Jiang and 4 other authors View PDF Abstract:Creating an immersive and interactive theatrical experience is a long-term goal in the field of interactive narrative. The emergence of large language models (LLMs) provides a new path to achieve this goal. However, existing LLM-based drama generation methods often produce models that lack initiative and cannot interact with the physical scene, while typically requiring detailed user input that diminishes the immersion of live performance. To address these challenges, we propose HAMLET, a hierarchical adaptive multi-agent framework focused on drama creation and real-time online performance. Given a simple topic, the framework first generates a narrative blueprint to guide the subsequent improvisational performance. In the online performance phase, each actor is equipped with an adaptive reasoning module that enables decision-making based on their personas, memories, goals, and emotional states during complex group chat scenarios. Beyond dialogue, actor agents engage in embodied interactions by changing t...