An Interpretable Memory Decision Controller for LLM Agents Based on Three-Signal Complementarity: Decoupling Confidence and Consistency
作者: Yiming Zhang, Jinghong Zhang, Haoran Zhao, Yiren Ma, Chunlei Zhao
分类: cs.CL
发布日期: 2026-09-18
备注: 17 pages, 6 figures, 10 tables
💡 一句话要点
提出可解释的记忆决策控制器以解决大语言模型的信任问题
🎯 匹配领域: 支柱九:具身大模型 (Embodied Foundation Models)
关键词: 记忆决策 大语言模型 可解释性 幻觉现象 风险评估 自然语言处理 智能助手
📋 核心要点
- 现有方法在处理冲突记忆时容易导致幻觉现象,缺乏对记忆可信度的有效评估。
- 本文提出的记忆决策层(MDL)通过三信号互补编码器,解耦信心与一致性,增强决策的可解释性。
- 实验表明,MDL显著降低了冲突记忆下的幻觉率,且在高风险场景中表现出色。
📝 摘要(中文)
大语言模型的记忆系统主要关注高效检索,而对检索记忆的信任决策关注较少。当记忆存储中存在冲突时,标准的检索增强生成方法会盲目注入记忆,从而加剧幻觉现象。本文提出了记忆决策层(MDL),作为检索与生成阶段之间的零参数控制器,通过三信号互补编码器融合相关性、可靠性和任务风险,生成可解释的决策表示,量化检索记忆的可信度。实验结果表明,MDL在冲突记忆下的幻觉率降低约56.04%,在高风险场景中接近零幻觉。
🔬 方法详解
问题定义:本文旨在解决大语言模型在面对冲突记忆时的信任决策问题。现有的检索增强生成方法未能有效评估记忆的可信度,导致幻觉现象加剧。
核心思路:提出的记忆决策层(MDL)通过三信号互补编码器,将相关性、可靠性和任务风险融合,生成可解释的决策表示,从而量化记忆的可信度。
技术框架:MDL位于检索与生成阶段之间,包含三信号互补编码器和风险反转机制,能够在决策过程中显式地解耦信心与一致性。
关键创新:MDL的主要创新在于其零参数设计和几何操作的使用,使其在速度上远超传统方法,并且提供了可解释的决策依据。
关键设计:MDL依赖于QR基的正交子空间投影和元工作记忆信号,确保决策过程的高效性和准确性,且每次决策仅增加约0.14毫秒的延迟。
🖼️ 关键图片
📊 实验亮点
实验结果显示,MDL在冲突记忆下的幻觉率降低约56.04%,在高风险场景中接近零幻觉,表现优于无记忆基线。该控制器的决策速度约为嵌入检索步骤的50倍,极大提升了系统的响应效率。
🎯 应用场景
该研究的潜在应用领域包括智能助手、自动问答系统和其他需要高可信度记忆检索的自然语言处理任务。通过提高模型对记忆的信任评估,能够显著提升用户体验和系统的可靠性,未来可能在医疗、法律等高风险领域发挥重要作用。
📄 摘要(原文)
Memory systems for large language models have focused predominantly on efficient retrieval, whereas the decision of whether retrieved memories should be trusted has received comparatively little attention. When the memory store contains conflicting positions, standard retrieval-augmented generation (RAG) blindly injects memories and amplifies hallucinations: in models susceptible to memory injection, the RAG hallucination rate under conflicting memories is markedly higher than that of a memory-free baseline. Inspired by memory signaling mechanisms in the prefrontal cortex, we propose the Memory Decision Layer (MDL), a zero-parameter memory decision controller situated between the retrieval and generation stages. Its core is a three-signal complementary encoder that fuses relevance, reliability, and task risk through QR-based orthogonal subspace projection and a meta-working-memory signal into an interpretable decision representation that quantifies the trustworthiness of retrieved memories. Building on this encoder, MDL explicitly decouples confidence from consistency and introduces risk inversion and explicit abstention. Evaluations on mainstream large language models and multiple open-source datasets show that MDL reduces the hallucination rate under conflicting memories by about 56.04% in general scenarios and approaches zero hallucination in high-risk scenarios. The controller is fully white-box: it relies purely on geometric operations, requires no trained parameters, and adds only about 0.14 ms per decision -- roughly 50x faster than the embedding-retrieval step that precedes it and four to five orders of magnitude faster than an LLM self-evaluation call.