HaReCAP: Habitual-action Grounding for Recursive Large Language Model Agents
作者: Shen Liu, Zhenguo Xu, Shaopu Wang, Yike Gao, Chunlei Wang
分类: cs.AI, cs.RO
发布日期: 2026-08-17
备注: 15 pages, 3 figures
💡 一句话要点
提出HaReCAP以解决长时间任务中的最后一步动作基础问题
🎯 匹配领域: 支柱九:具身大模型 (Embodied Foundation Models)
关键词: 长时间任务 大型语言模型 递归上下文管理 动作基础 机器人导航 智能代理 效率提升
📋 核心要点
- 现有的递归上下文管理方法在长时间任务中存在最后一步动作基础冗余的问题,导致LLM调用频繁且效率低下。
- HaReCAP通过提取成功轨迹中的常见决策,离线生成可审计的叶子反应规则,从而减少LLM的调用频率。
- 在Robotouille和ALFWorld的实验中,HaReCAP分别减少了14.67%、17.93%和20.08%的token消耗,显示出显著的性能提升。
📝 摘要(中文)
长时间的具身任务要求大型语言模型(LLM)代理逐步分解高层目标,并根据环境反馈修订计划。现有的递归上下文管理方法ReCAP在叶子节点上多次调用LLM以将原子子任务转化为可执行动作,导致了最后一步基础冗余问题。为此,本文提出了HaReCAP(习惯性动作基础ReCAP),通过提取成功轨迹中的频繁叶子决策,离线编译为可审计的单步叶子反应规则,从而减少LLM调用次数。在Robotouille和ALFWorld上的实验表明,HaReCAP在多个任务中显著降低了token消耗,证明了其作为ReCAP风格递归上下文管理框架的低侵入性扩展的有效性。
🔬 方法详解
问题定义:本文旨在解决长时间具身任务中LLM代理在叶子节点上频繁调用导致的基础冗余问题,造成了不必要的token消耗和计算开销。
核心思路:HaReCAP通过提取成功轨迹中的频繁决策,生成可审计的单步叶子反应规则,允许在特定情况下跳过LLM调用,从而提高效率。
技术框架:HaReCAP的整体架构包括离线规则生成和在线决策过程。离线阶段提取成功轨迹中的决策并生成规则,在线阶段根据当前有效动作集判断是否使用规则。
关键创新:HaReCAP的主要创新在于引入了可审计的叶子反应规则,允许在特定情况下避免LLM调用,这与传统的ReCAP方法形成了鲜明对比。
关键设计:在设计中,HaReCAP的规则生成过程依赖于成功轨迹的分析,确保生成的规则能够唯一确定当前有效动作,同时保持原有的递归控制流程。
🖼️ 关键图片
📊 实验亮点
HaReCAP在Robotouille和ALFWorld的实验中表现出色,分别减少了14.67%、17.93%和20.08%的token消耗,证明了其在长时间任务中的有效性和优势,显著提升了任务执行的效率。
🎯 应用场景
HaReCAP的研究成果在机器人导航、智能家居和自动化任务等领域具有广泛的应用潜力。通过减少LLM调用频率,能够显著提高系统的响应速度和效率,推动智能代理在复杂环境中的应用。
📄 摘要(原文)
Long-horizon embodied tasks require LLM agents to iteratively decompose high-level goals, revise plans in response to environmental feedback, and ground leaf-level subgoals into valid executable actions. Recursive context-management methods such as ReCAP improve planning stability through multi-level task decomposition and parent-node refinement, but still repeatedly invoke the LLM at leaf nodes to ground atomic subtasks into exact valid actions. We refer to this final grounding step as last-mile grounding redundancy, which accumulates into substantial LLM-call and token overhead during long-horizon execution. To mitigate this issue, we propose HaReCAP (Habitual-action Grounded ReCAP), a low-intrusion leaf grounding extension for ReCAP. HaReCAP extracts frequent leaf decisions from successful trajectories and compiles them offline into auditable and abstainable one-step leaf-reflex rules. At runtime, it skips the leaf LLM call only when a rule can uniquely determine a legal action in the current valid-action set; otherwise, it falls back to the original ReCAP. This design avoids repeatedly carrying the full recursive context into the LLM for routine leaf action grounding, while preserving the original recursive control flow. We evaluate HaReCAP on Robotouille and ALFWorld with Qwen3.5-27B as the main model. On tasks solved by both ReCAP and HaReCAP, HaReCAP reduces token consumption by 14.67%, 17.93%, and 20.08% on Robotouille synchronous, Robotouille asynchronous, and ALFWorld, respectively. The results show that HaReCAP can serve as a low-intrusion extension to ReCAP-style recursive context-management frameworks, reducing last-mile grounding redundancy across environments and models on commonly successful trajectories.