CovNav: Conformal Prediction-Guided Adaptive Inference for Object Goal Navigation

📄 arXiv: 2610.27677v1 📥 PDF

作者: Pengying Wu, Zhejia Wen, Kang Ding, Kangjie Zhou, Chunxuan Jiao, Wei Pan, Xiangli Nie, Chang Liu

分类: cs.RO, cs.AI, cs.CL, cs.CV

发布日期: 2026-10-20

备注: 13 pages, 6 figures, 7 tables

💡 一句话要点

提出CovNav以解决物体目标导航中的推理效率问题

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

关键词: 物体目标导航 保形预测 具身智能体 推理效率 大型语言模型 空间常识 慢思考推理 不确定性感知

📋 核心要点

  1. 现有方法在物体目标导航中面临推理质量与效率之间的权衡，尤其是在使用大型语言模型时。
  2. CovNav通过保形预测实现了不确定性感知决策，将双过程推理转化为结构化的二元决策，提升了推理效率。
  3. 实验结果显示，CovNav在HM3D v1和v2上成功率分别提高0.8和1.7个百分点，同时仅在12.4%的查询中使用慢推理。

📝 摘要（中文）

在物体目标导航中，具身智能体需要在未见环境中找到目标物体，依赖空间常识。现有的大型语言模型在推理质量与效率之间存在权衡。本文提出CovNav，一个无训练框架，利用保形预测将双过程推理转化为不确定性感知决策。CovNav通过快速模式判断目标物体是否靠近观察到的物体，并将模糊查询升级到慢思考推理。实验表明，CovNav在多个基准上超越了最强的零-shot基线，同时仅在12.4%的查询中调用慢推理，显著提高了推理效率。

🔬 方法详解

问题定义：本文解决的是物体目标导航中具身智能体如何在未见环境中高效找到目标物体的问题。现有方法在推理时面临质量与效率的权衡，尤其是大型语言模型的使用带来了延迟和隐私风险。

核心思路：CovNav的核心思路是利用保形预测将双过程推理转化为不确定性感知的决策过程。通过快速模式判断目标物体的可能位置，并仅在不确定时调用慢思考推理，从而提高推理效率。

技术框架：CovNav的整体架构包括一个快速模式的本地语言模型和保形预测模块。快速模式负责初步判断目标物体的可能性，而保形预测则将其概率转化为预测集，模糊查询才会被升级到慢思考推理。

关键创新：CovNav的主要创新在于将空间查询视为结构化的二元决策，而非自由形式生成。这种方法使得不确定性感知决策成为可能，显著提升了推理效率。

关键设计：CovNav在设计上采用了本地语言模型进行快速判断，并通过保形预测将模型输出的概率转化为预测集。此外，采用了现成的检测器来支持开放集物体的识别，保持了校准的可操作性。

🖼️ 关键图片

fig_0
fig_1
fig_2

📊 实验亮点

CovNav在HM3D v1和v2上的成功率分别提高了0.8和1.7个百分点，且仅在12.4%的查询中调用慢推理。这表明CovNav在保持高效推理的同时，能够有效利用慢思考推理的优势，显著提升了整体性能。

🎯 应用场景

CovNav的研究成果在机器人导航、智能家居和自动化服务等领域具有广泛的应用潜力。通过提高具身智能体在复杂环境中的导航能力，CovNav能够有效提升机器人在实际任务中的表现，推动智能机器人技术的进步。

📄 摘要（原文）

In object goal navigation (ObjectNav), an embodied agent must find a target object (e.g., "find a bed") in a previously unseen environment, relying on spatial commonsense such as object co-occurrence. Large Language Models (LLMs) encode this commonsense, but deploying them on robots exposes a trade-off between reasoning quality and efficiency: cloud-based models incur latency and privacy risks, whereas small on-device LLMs are fast but prone to hallucinated spatial judgments. Hybrid reasoning LLMs offer both a fast non-thinking mode and a slow thinking mode, which raises a key question for embodied agents: when is slow thinking worth its cost? We propose CovNav, a training-free framework that answers this question with Conformal Prediction (CP), turning dual-process (System 1 / System 2) inference into an uncertainty-aware decision. Much like recent decision models such as Jev, CovNav casts each spatial query as a structured binary decision rather than free-form generation: a local LLM in fast mode judges whether the target is likely near an observed object, CP converts its token probabilities into a prediction set, and only ambiguous queries are escalated to slow chain-of-thought reasoning. The object-pair formulation keeps calibration tractable and supports open-set objects through off-the-shelf detectors. Experiments on Gibson and HM3D show that CovNav surpasses the strongest zero-shot baseline by 0.8 and 1.7 percentage points in success rate on HM3D v1 and v2, while invoking slow reasoning for only 12.4% of queries and retaining most of the gain of always-on thinking at a fraction of its latency. Real-robot experiments on a TurtleBot 4 further validate the approach. CovNav shows how calibrated uncertainty can decide when an embodied agent should think slowly, offering a principled form of adaptive test-time reasoning for LLM-based robots.