外观
论文arXiv 2509.226432025-09-26
VLA-Reasoner：MCTS赋能视觉语言动作模型长程操作推理
本文由 AI 根据公开内容整理，可能有误，请以原文为准：VLA-Reasoner: Empowering Vision-Language-Action Models with Reasoning via Online Monte Carlo Tree Search（arXiv 2509.22643）
核心内容
针对现有视觉语言动作（VLA）模型仅能预测短时动作、在长程轨迹任务中易产生累积偏差的问题，研究团队提出即插即用框架VLA-Reasoner。该方法不改变基座模型权重，而是通过测试时计算扩展（Test-time Scaling）赋予现成VLA预见未来状态的能力。实验表明，该框架在仿真与真机环境中均显著优于当前SOTA VLA模型，为机器人操作提供了可扩展的推理新范式。
技术上，VLA-Reasoner利用世界模型对动作轨迹进行Rollout生成未来状态作为推理依据，并采用蒙特卡洛树搜索（MCTS）在大动作空间中高效寻优。系统以逐步VLA预测作为搜索根节点，引入基于核密度估计（KDE）的置信采样机制避免冗余查询，同时通过离线价值估计策略评估中间状态得分，利用长期反馈修正执行偏差，实现了从“反应式”到“规划式”操作的跨越。
关键看点
- 测试时推理：无需微调基座VLA，通过在线MCTS与世界模型在推理阶段动态规划最优动作序列
- 高效搜索机制：采用基于核密度估计的置信采样策略，在减少冗余VLA查询的同时保障大空间探索效率
- 长程纠偏能力：引入离线价值估计策略对MCTS中间状态评分，利用长期反馈修正增量偏差解决短视问题
- 通用增强范式：作为插件框架兼容现有VLA，在仿真与真机长程操作任务中验证了性能显著提升
工程与产业观察
该工作证实了“系统2”慢思考在具身智能中的工程价值。对于算法工程师而言，在不重训大模型的前提下，通过外挂世界模型与搜索算法扩展测试时算力，是解决VLA长程规划短板的高性价比路径。这标志着具身智能正从单纯的模仿学习向“感知-规划-决策”闭环演进，为复杂工业与服务场景落地提供了关键技术支撑。
英文摘要 (Abstract)
Vision-Language-Action models (VLAs) achieve strong performance in general robotic manipulation tasks by scaling imitation learning. However, existing VLAs are limited to predicting short-sighted next-action, which struggle with long-horizon trajectory tasks due to incremental deviations. To address this problem, we propose a plug-in framework named \method that effectively empowers off-the-shelf VLAs with the capability of foreseeing future states via test-time scaling. Specifically, \method samples and rolls out possible action trajectories where involved actions are rationales to generate future states via a world model, which enables \method to foresee and reason potential outcomes and search for the optimal actions. We further leverage Monte Carlo Tree Search (MCTS) to improve search efficiency in large action spaces, where step-wise VLA predictions seed the root. Meanwhile, we introduce a confidence sampling mechanism based on Kernel Density Estimation (KDE), to enable efficient exploration in MCTS without redundant VLA queries. We evaluate intermediate states in MCTS via an offline value estimation strategy, to score predicted futures and correct deviations with long-term feedback. We conducted extensive experiments in both simulators and the real world, demonstrating that our proposed VLA-Reasoner achieves significant improvements over the state-of-the-art VLAs. Our method highlights a potential pathway toward scalable test-time computation of robotic manipulation. The project website is available at: https://vla-reasoner.github.io/.
作者：Wenkai Guo, Guanxing Lu, Haoyuan Deng, Zhenyu Wu, Yansong Tang, Ziwei Wang
来源
本页事实对应的公开页面，以原文为准。