From Reasoning to Action: Consequence-Grounded Alignment for Vision-Language-Action Models

📄 arXiv: 2610.39543v1 📥 PDF

作者: Yixue Zhang, Wenke Xia, Kun Wu, Wenbo Yu, Pei Ren, Junnan Nie, Di Hu, Zhengping Che, Chang Liu

分类: cs.RO

发布日期: 2026-10-29

备注: Project page: https://onlyxingxingzyx.github.io/CoRAct/

🔗 代码/项目: PROJECT_PAGE

💡 一句话要点

提出CoRAct框架以解决视觉-语言-行动模型的对齐问题

🎯 匹配领域: 支柱一：机器人控制 (Robot Control) 支柱九：具身大模型 (Embodied Foundation Models)

关键词: 视觉-语言-行动 推理对齐 奖励引导优化 具身思维链 机器人操作

📋 核心要点

  1. 现有的视觉-语言-行动模型在推理与行动之间的对齐上存在隐式性，导致生成的行动与预期后果不一致。
  2. CoRAct框架通过奖励引导优化，明确将推理与行动对齐，利用运动和航点作为可验证信号。
  3. 在RoboCasa、VLABench和真实操作任务中的实验结果显示，CoRAct在任务成功率和推理执行对齐上显著优于现有基线。

📝 摘要（中文）

推理通过将预期行动与其后果连接起来来指导决策。近期的视觉-语言-行动（VLA）模型采用了具身思维链（ECoTs）来引导行动生成，但这种引导仍然是隐式的。本文提出了CoRAct，一个基于后果的推理到行动对齐框架，通过在行动噪声空间中进行奖励引导优化，明确对齐推理与行动。CoRAct将两个物理基础的ECoT组件定义为可验证信号：运动指定预期运动，而航点编码预期空间后果。实验表明，CoRAct在任务成功率、泛化能力和推理-执行对齐方面相较于强基线有显著提升。

🔬 方法详解

问题定义：本文旨在解决现有视觉-语言-行动模型中推理与行动之间的隐式对齐问题，导致生成的行动与预期后果不一致。

核心思路：CoRAct框架通过奖励引导优化，明确对齐推理与行动，利用运动和航点作为可验证信号，从而提高行动生成的准确性和一致性。

技术框架：CoRAct的整体架构包括两个主要模块：物理基础的ECoT组件（运动和航点）和一个统一的混合变换器模型，后者共同建模ECoT推理、行动生成和未来预测。

关键创新：CoRAct的主要创新在于其奖励引导优化机制，通过在行动噪声空间中计算对齐奖励，明确优化初始噪声分布，从而提高行动生成的质量。

关键设计：在设计中，CoRAct使用运动一致性和预测结果与航点之间的协议来计算对齐奖励，确保生成的行动与预期后果相一致。

🖼️ 关键图片

fig_0
fig_1
fig_2

📊 实验亮点

实验结果显示，CoRAct在RoboCasa、VLABench和真实操作任务中，任务成功率提高了20%以上，推理与执行的对齐度显著增强，相较于强基线模型表现出更好的泛化能力。

🎯 应用场景

该研究在机器人操作、智能家居和人机交互等领域具有广泛的应用潜力。通过将推理与行动明确对齐，CoRAct能够在新场景中实现零样本适应，提升系统的灵活性和智能化水平。

📄 摘要（原文）

Reasoning informs decision-making by connecting intended actions with their expected consequences. Recent vision-language-action (VLA) models incorporate embodied chains of thought (ECoTs) to guide action generation. However, such guidance remains implicit, as actions are conditioned on ECoTs rather than explicitly optimized to align with them. We introduce CoRAct, a consequence-grounded reasoning-to-action alignment framework that explicitly aligns reasoning and action through reward-guided optimization in action noise space. CoRAct formulates two physically grounded ECoT components as verifiable signals: movement specifies intended motion, while waypoints encode expected spatial consequences. A unified Mixture-of-Transformers jointly models ECoT reasoning, action generation, and action-conditioned future prediction. For candidate actions generated from sampled noise, CoRAct computes alignment rewards based on movement consistency and agreement between predicted outcomes and specified waypoints. These rewards guide group-relative optimization of the initial noise distribution, steering action generation toward consistency with intended actions and expected consequences. Experiments across RoboCasa, VLABench, and real-world manipulation tasks demonstrate improvements in task success, generalization, and reasoning-execution alignment over strong ECoT-based VLA baselines. CoRAct also grounds externally provided ECoTs in executable behavior, enabling zero-shot adaptation to novel scenarios by leveraging knowledge from powerful vision-language models. Project page and dataset: https://onlyxingxingzyx.github.io/CoRAct/