DREAM: LLM-based Dynamic Role-playing via Event-Aware Memory Graph
This work addresses the problem of maintaining long-term narrative and personality coherence in role-playing agents, offering a structured memory approach that improves interpretability and consistency for developers and users of such systems.
DREAM introduces a structured memory framework for role-playing agents, using an Event-aware Memory Graph based on the ABC cognitive model to organize character experiences into temporally ordered and causally linked events. It achieves state-of-the-art performance on CoSER, LIFECHOICE, and a new TCM benchmark, outperforming multiple strong baselines.
Role-playing agents (RPAs) have emerged as a key application of large language models, enabling immersive and high-fidelity character simulation. Accurate role-playing of established characters requires not only stylistic imitation but also temporally consistent and causally grounded behavioral reasoning. However, existing RPAs primarily rely on static character descriptions and unstructured memory, limiting their ability to maintain long-term narrative and personality coherence. We introduce DREAM, a structured memory framework for role-playing agents inspired by the Activating Event-Belief-Consequence (ABC) cognitive model. DREAM transforms unstructured literary text into an Event-aware Memory Graph (EMG) that organizes character experiences into temporally ordered and causally linked event graph. This representation enables the construction of dynamic, dual-granularity character profiles that capture both stable personality traits and event-driven behavioral evolution. We further propose the Temporal Causal Memory (TCM) benchmark to evaluate temporal consistency and long-range causal narrative coherence. DREAM achieves state-of-the-art performance across CoSER, LIFECHOICE, and TCM, outperforming multiple strong baselines. Our approach demonstrates the effectiveness of structured memory in enhancing the interpretability and consistency of role-playing agents.