从历史到未来：以经验与思考演化智能体实现零样本视觉语言导航
History to Future: Evolving Agent with Experience and Thought for Zero-shot Vision-and-Language Navigation
把 LLM 的未来动作/地标预测（F-CoT）与历史轨迹经验总结（H-CoE）组合成可演化的零样本导航决策回路，可直接迁移到视觉导航、具身探索等需要在线纠错的序列决策任务中。
原文依据 F-CoT predicts future actions and landmarks as thoughts to assist navigation progress estimation and direction selection
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论连续环境下的视觉语言导航（VLN-CE）要求智能体遵循语言指令导航至目标地点。随着大语言模型（LLM）的发展，近期工作探索将其适配于零样本 VLN-CE，为解决基于训练的范式泛化能力差的缺陷提供了有前景的方案。然而，现有基于 LLM 的工作主要进行朴素推理以做决策，且缺乏反馈，例如回顾历史错误与预测未来潜力。因此，它可能在那些初始即出错的任务上持续失败。本文重新审视基于 LLM 的零样本 VLN-CE，并提出一种新范式 EvoNav，通过未来思维链（F-CoT）与历史经验链（H-CoE）以未来思考与历史经验改进智能体的决策。F-CoT 预测未来动作与地标作为思考，以辅助导航进度估计与方向选择；H-CoE 则将历史轨迹与场景总结为经验，以提升导航决策的可靠性。F-CoT 与 H-CoE 协同演化智能体的决策。在仿真器与真实环境中的大量实验证明了 EvoNav 的有效性。
英文原文摘要 1321 字符
Vision-and-Language Navigation in Continuous Environment (VLN-CE) requires an agent to follow language instructions to navigate the target destination. With the advancement of large language models (LLMs), recent efforts have explored adapting them for zero-shot VLN-CE, offering a promising solution in addressing the drawbacks of poor generalization in the training-based paradigm. However, existing LLM-based works primarily perform naive reasoning for decision-making and lack feedback, e.g., reviewing historical errors and predicting future potentials. Consequently, it may suffer from continuous failure for those initial error tasks. In this paper, we rethink LLM-based zero-shot VLN-CE and propose a new paradigm, named EvoNav, to improve the agent's decision-making with future thought and history experience via Future Chain-of-Thought (F-CoT) and History Chain-of-Experience (H-CoE). F-CoT predicts future actions and landmarks as thoughts to assist navigation progress estimation and direction selection, while H-CoE summarizes historical trajectories and scenes as experience to improve navigation decision reliability. Both F-CoT and H-CoE cooperatively evolve the agent's decision-making. Extensive experiments in both the simulator and real-world environments demonstrate the effectiveness of our EvoNav.