让机器人行动与思考对齐，提升复杂任务的鲁棒性。
Do What You Say: Steering Vision-Language-Action Models via Runtime Reasoning-Action Alignment Verification
- 用模拟预测多种动作结果，选最符合文本计划的执行
- 在分布外场景下性能提升15%，支持新行为组合无需重训
- 无需训练，可直接增强现有推理型视觉语言动作模型
推理型视觉语言动作（VLA）模型通过生成分步文本计划来提升机器人指令执行能力，受语言模型链式思维（CoT）启发。然而即使文本计划正确，生成的动作仍可能偏离预期目标，尤其在分布外（OOD）场景中。本文将此现象定义为具身链式思维忠实性不足，并提出一种无需训练的运行时策略调优方法。给定VLA的中间文本计划，框架从同一模型采样多个候选动作序列，通过仿真预测其结果，并使用预训练视觉语言模型（VLM）选择与文本计划最匹配的动作序列。仅执行与文本推理一致的动作，将原始VLA的动作多样性由误差源转化为优势，显著提升对语义和视觉分布外扰动的鲁棒性，并实现无需代价重训的新行为组合。我们还构建了带推理标注的LIBERO-100扩展数据集，设计针对OOD评估的环境变体，在行为组合任务上相比先前工作最高提升15%，且性能随计算资源与数据多样性增长。项目网站：https://yilin-wu98.github.io/steering-reasoning-vla/
原文摘要 · Abstract (English)
Reasoning Vision Language Action (VLA) models improve robotic instruction-following by generating step-by-step textual plans before low-level actions, an approach inspired by Chain-of-Thought (CoT) reasoning in language models. Yet even with a correct textual plan, the generated actions can still miss the intended outcomes in the plan, especially in out-of-distribution (OOD) scenarios. We formalize this phenomenon as a lack of embodied CoT faithfulness, and introduce a training-free, runtime policy steering method for reasoning-action alignment. Given a reasoning VLA's intermediate textual plan, our framework samples multiple candidate action sequences from the same model, predicts their outcomes via simulation, and uses a pre-trained Vision-Language Model (VLM) to select the sequence whose outcome best aligns with the VLA's own textual plan. Only executing action sequences that align with the textual reasoning turns our base VLA's natural action diversity from a source of error into a strength, boosting robustness to semantic and visual OOD perturbations and enabling novel behavior composition without costly re-training. We also contribute a reasoning-annotated extension of LIBERO-100, environment variations tailored for OOD evaluation, and demonstrate up to 15% performance gain over prior work on behavior composition tasks and scales with compute and data diversity. Project Website at: https://yilin-wu98.github.io/steering-reasoning-vla/
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。