为智能体设计可记忆人物与场景的事件型记忆系统
CAST: Character-and-Scene Episodic Memory for Agents
- 用人物与时空场景构建3维事件记忆结构
- 在对话任务上平均提升8.11%准确率和10.21%评分
- 适合需要理解时间线与角色互动的应用
情景记忆是人类记忆的核心，指对包含人物、时间和地点的连贯事件的回忆能力。然而，多数智能体记忆系统仅关注语义召回，将经验视为键值对、向量或图结构，难以表征和检索连贯事件。为此，我们提出受戏剧理论启发的基于人物与场景的记忆架构（CAST）。CAST 构建三维场景（时间/地点/主题），并将它们组织成描述人物行为的档案以表示情景记忆。此外，CAST 还结合基于图的语义记忆，形成稳健的双重记忆设计。实验表明，在多个数据集上，CAST 在各类对话任务中平均提升 8.11% F1 分数和 10.21% J（LLM-as-a-Judge）得分，尤其在开放性和时间敏感性对话问题上表现突出。
原文摘要 · Abstract (English)
Episodic memory is a central component of human memory, which refers to the ability to recall coherent events grounded in who, when, and where. However, most agent memory systems only emphasize semantic recall and treat experience as structures such as key-value, vector, or graph, which makes them struggle to represent and retrieve coherent events. To address this challenge, we propose a Character-and-Scene based memory architecture(CAST) inspired by dramatic theory. Specifically, CAST constructs 3D scenes (time/place/topic) and organizes them into character profiles that summarize the events of a character to represent episodic memory. Moreover, CAST complements this episodic memory with a graph-based semantic memory, which yields a robust dual memory design. Experiments demonstrate that CAST has averagely improved 8.11% F1 and 10.21% J(LLM-as-a-Judge) than baselines on various datasets, especially on open and time-sensitive conversational questions.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。