基于互联网规模知识的自监督行动预测具身推理
Self-Supervised Bootstrapping of Action-Predictive Embodied Reasoning
AI总结：
本文提出R&B-EnCoRe方法，通过自监督细化使模型从互联网知识中自推导具身推理策略，提升动作执行和导航性能，减少碰撞率。
摘要：
具身链式思维（CoT）推理显著提升了视觉-语言-动作（VLA）模型，但当前方法依赖刚性模板指定推理原语（如场景中的物体、高层计划、结构 affordances）。这些模板可能迫使策略处理无关信息，干扰关键动作预测信号。我们引入R&B-EnCoRe，使模型通过自监督细化从互联网规模知识中自推导具身推理。通过将推理视为重要加权变分推断中的潜在变量，模型可生成并提炼无外部奖励、验证者或人工标注的具身特定策略训练数据集。我们在各种VLA架构中验证R&B-EnCoRe，应用于 manipulation（Franka Panda在仿真中，WidowX在硬件中）、legged导航（双足、轮式、自行车、四足）和自动驾驶具身，参数规模为1B、4B、7B和30B。我们的方法在 manipulation 成功率提升28%，导航评分提高101%，碰撞率减少21%。R&B-EnCoRe使模型提炼出预测成功控制的推理，避免手动标注工程，同时将互联网规模知识接地于物理执行。
英文摘要：
Embodied Chain-of-Thought (CoT) reasoning has significantly enhanced Vision-Language-Action (VLA) models, yet current methods rely on rigid templates to specify reasoning primitives (e.g., objects in the scene, high-level plans, structural affordances). These templates can force policies to process irrelevant information that distracts from critical action-prediction signals. This creates a bottleneck: without successful policies, we cannot verify reasoning quality; without quality reasoning, we cannot build robust policies. We introduce R&B-EnCoRe, which enables models to bootstrap embodied reasoning from internet-scale knowledge through self-supervised refinement. By treating reasoning as a latent variable within importance-weighted variational inference, models can generate and distill a refined reasoning training dataset of embodiment-specific strategies without external rewards, verifiers, or human annotation. We validate R&B-EnCoRe across manipulation (Franka Panda in simulation, WidowX in hardware), legged navigation (bipedal, wheeled, bicycle, quadruped), and autonomous driving embodiments using various VLA architectures with 1B, 4B, 7B, and 30B parameters. Our approach achieves 28% gains in manipulation success, 101% improvement in navigation scores, and 21% reduction in collision-rate metric over models that indiscriminately reason about all available primitives. R&B-EnCoRe enables models to distill reasoning that is predictive of successful control, bypassing manual annotation engineering while grounding internet-scale knowledge in physical execution.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
