探索性与同化性反思：用于长期记忆的反思性回忆循环
Exploratory and Assimilating Reflection: Reflective Recall Cycle for Long-term Memory
AI总结：
研究基于大语言模型的自主智能体外部记忆问题，提出探索性 - 同化性反思（EAR）框架，结合探索性反思与同化性反思两种机制，提高初始检索性能和样本效率，在长期对话基准上比基线检索器提升检索率达 17.9%，且样本效率高、抗噪声反馈。
AI中文摘要：
基于大语言模型的自主智能体需要外部记忆来克服其无状态性以及长期交互和动态知识推理中的有限上下文窗口问题。然而，现有记忆检索方法往往缺乏适应性和样本效率，难以从异构存储中检索出合适的记忆组合。我们提出了探索性 - 同化性反思（EAR）框架，用于实现高初始检索性能和样本高效适应。EAR 结合了两种机制：探索性反思，通过迭代搜索引导检索并为每个查询收集有用经验；同化性反思，从经验缓冲区重放这些经验，比仅依赖即时奖励的方法更有效地优化全局重排器。实验表明，EAR 在两个长期对话基准上比基线检索器提高了高达 17.9%的检索率。我们还表明 EAR 具有高度样本效率且对噪声反馈具有鲁棒性。
英文摘要：
LLM-based autonomous agents require external memory to overcome their statelessness and limited context window for long-term interaction and dynamic knowledge reasoning. However, existing memory retrieval methods often lack adaptability and sample efficiency, and struggle to retrieve the right mixture of memories from heterogeneous stores. We propose Exploratory-Assimilating Reflection (EAR), a framework for high initial retrieval performance and sample-efficient adaptation. EAR combines two mechanisms: Exploratory Reflection, which performs iterative search to bootstrap retrieval and collect useful experiences for each query, and Assimilating Reflection, which replays these experiences from an Experience Buffer to refine a global reranker more efficiently than methods relying only on immediate rewards. Experiments show that EAR improves retrieval by up to 17.9% over the baseline retriever on two long-term dialogue benchmarks. We also show that EAR is highly sample-efficient and robust to noisy feedback.
