Tree-Planner: Efficient Close-loop Task Planning with Large Language Models

📄 arXiv: 2310.08582v2 📥 PDF

作者: Mengkang Hu, Yao Mu, Xinmiao Yu, Mingyu Ding, Shiguang Wu, Wenqi Shao, Qiguang Chen, Bin Wang, Yu Qiao, Ping Luo

分类: cs.CL, cs.AI, cs.LG, cs.RO

发布日期: 2023-10-12 (更新: 2024-07-24)

备注: Published in ICLR 2024

💡 一句话要点

提出Tree-Planner以解决大语言模型任务规划中的效率问题

🎯 匹配领域: 支柱九：具身大模型 (Embodied Foundation Models)

关键词: 闭环任务规划 大语言模型 动作树 计划采样 实时决策 效率优化 错误修正

📋 核心要点

  1. 现有方法在使用大语言模型进行任务规划时，存在高令牌消耗和冗余错误修正的问题，影响了其可扩展性。
  2. 论文提出Tree-Planner，通过计划采样、动作树构建和基于环境信息的决策三个阶段来优化任务规划过程。
  3. 实验结果显示，Tree-Planner在效率和性能上均优于现有方法，显著降低了令牌消耗和错误修正次数。

📝 摘要（中文）

本文研究了闭环任务规划，即在特定目标下生成技能序列（计划），并根据实时观察调整计划。近年来，利用大语言模型（LLMs）迭代生成动作已成为一种流行的范式，但面临高令牌消耗和冗余错误修正等低效问题。为此，本文提出了Tree-Planner，将任务规划分为三个阶段：计划采样、动作树构建和基于环境信息的决策。实验表明，Tree-Planner在保持高效率的同时实现了最先进的性能，相较于之前的最佳模型，令牌消耗减少了92.2%，错误修正减少了40.5%。

🔬 方法详解

问题定义：本文旨在解决在闭环任务规划中，现有大语言模型方法的高令牌消耗和冗余错误修正问题，这些问题限制了其在大规模应用中的可行性。

核心思路：Tree-Planner通过将任务规划过程分为三个阶段，优化了大语言模型的使用效率。首先进行计划采样，然后构建动作树，最后基于实时环境信息进行决策，从而减少了重复令牌消耗和错误修正。

技术框架：Tree-Planner的整体架构包括三个主要模块：计划采样模块，利用LLM生成潜在计划；动作树构建模块，将多个计划聚合成树结构；决策模块，基于环境信息进行自上而下的决策。

关键创新：最重要的创新在于将任务规划过程重构为三个阶段，特别是通过动作树的构建和决策过程，显著提高了效率和灵活性，减少了冗余的令牌消耗。

关键设计：在设计中，Tree-Planner通过将LLM查询分解为单个计划采样调用和多个基于环境的决策调用，显著降低了令牌消耗。此外，支持在动作树上进行回溯，使得错误修正过程更加灵活。

🖼️ 关键图片

fig_0
fig_1
fig_2

📊 实验亮点

实验结果显示，Tree-Planner在效率上表现优异，相较于之前的最佳模型，令牌消耗减少了92.2%，同时错误修正次数减少了40.5%，展现出显著的性能提升。

🎯 应用场景

该研究的潜在应用领域包括机器人控制、智能助手和自动化系统等，能够在动态环境中实现高效的任务规划。Tree-Planner的设计理念和方法可以为未来的智能系统提供更高效的决策支持，推动相关技术的发展与应用。

📄 摘要（原文）

This paper studies close-loop task planning, which refers to the process of generating a sequence of skills (a plan) to accomplish a specific goal while adapting the plan based on real-time observations. Recently, prompting Large Language Models (LLMs) to generate actions iteratively has become a prevalent paradigm due to its superior performance and user-friendliness. However, this paradigm is plagued by two inefficiencies: high token consumption and redundant error correction, both of which hinder its scalability for large-scale testing and applications. To address these issues, we propose Tree-Planner, which reframes task planning with LLMs into three distinct phases: plan sampling, action tree construction, and grounded deciding. Tree-Planner starts by using an LLM to sample a set of potential plans before execution, followed by the aggregation of them to form an action tree. Finally, the LLM performs a top-down decision-making process on the tree, taking into account real-time environmental information. Experiments show that Tree-Planner achieves state-of-the-art performance while maintaining high efficiency. By decomposing LLM queries into a single plan-sampling call and multiple grounded-deciding calls, a considerable part of the prompt are less likely to be repeatedly consumed. As a result, token consumption is reduced by 92.2% compared to the previously best-performing model. Additionally, by enabling backtracking on the action tree as needed, the correction process becomes more flexible, leading to a 40.5% decrease in error corrections.