发表机构
Keio University（庆应义塾大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
提出NarrativeFlow，一种基于流匹配的语言条件机器人速度场模型，解决多平台数据利用与语言操作难题，在标准数据集和真实实验中优于基线方法。
AI 中文摘要
我们聚焦于语言条件下的流式操作，其中机器人流（机器人速度场）作为与具体实体无关的、以运动为中心的表示，用于利用从多个机器人平台收集的数据。该任务至关重要，因为语言条件下的操作对于实际机器人系统必不可少，然而扩展机器人基础模型仍受限于特定实体数据收集的高劳动强度。现有方法要么用稀疏关键点位移粗略近似机器人流，要么无法处理语言条件下的操作。为解决此局限，我们提出NarrativeFlow，该模型利用以语言为条件的流匹配公式，将机器人流建模为连续速度场。据此，NarrativeFlow生成的机器人流与实际操作在物理上保持一致。为验证NarrativeFlow，我们在语言条件操作的标准数据集上进行了实验。实验结果表明，NarrativeFlow在标准评估指标上优于代表性基线方法。此外，通过真实世界实验，我们展示了NarrativeFlow在多个操作任务中比基线方法实现了更高的成功率。项目页面可从此URL获取。
英文摘要
We focus on language-conditioned flow-based manipulation, where robot flows (robot velocity fields) serve as embodiment-agnostic, motion-centric representations for leveraging data collected from multiple robot platforms. This task is crucial because language-conditioned manipulation is essential for practical robotic systems, yet scaling robot foundation models remains limited by the labor-intensive collection of embodiment-specific data. Existing methods either coarsely approximate robot flows with sparse keypoint displacements, or cannot handle language-conditioned manipulation. To address this limitation, we propose NarrativeFlow, which models robot flows as continuous velocity fields using a flow-matching formulation conditioned on language. Accordingly, NarrativeFlow generates robot flows that are physically consistent with real-world manipulation. To validate NarrativeFlow, we have conducted experiments on standard datasets for language-conditioned manipulation. The experimental results show that NarrativeFlow outperforms representative baseline methods on standard evaluation metrics. Furthermore, through real-world experiments, we show that NarrativeFlow achieves higher success rates than baseline methods across multiple manipulation tasks. The project page is available at https://shota0520.github.io/NarrativeFlow-project-page/
CommentsAccepted at ACCV 2026
