外观
论文HF Daily Papers · arXiv 2610.329655 天前
RynnUnify 16B统一具身模型实现视觉规划与机器人控制闭环
本文由 AI 根据公开内容整理，可能有误，请以原文为准：RynnUnify: Towards Unified Embodied Brain（HF Daily Papers · arXiv 2610.32965）
核心内容
研究团队发布16B参数统一具身模型RynnUnify，旨在解决具身智能中物理理解与长程视觉规划割裂的问题。该模型在多项评测中取得最高平均具身理解分数，长程视觉规划基准Rynn-VP总分较Cosmos 3 Nano提升9.75分，验证了其在空间推理、物理预测及接地能力上的综合优势，为统一具身大脑提供了新范式。
技术上采用交错视频-语言推理架构，并通过强化学习开发Agentic扩展以选择性利用视频预测推断未观测状态，使Rynn-GEU平均分提升6.0分。World-Action扩展将理解预测转化为控制指令，在20个RoboCasa365复合任务中成功率达51.5%，三项真机任务平均成功率46.7%，实现了从认知到执行的端到端闭环验证。
关键看点
- 统一架构：16B参数模型融合具身理解与长程视觉规划，Rynn-VP基准超Cosmos 3 Nano 9.75分
- 强化学习扩展：Agentic模块选择性调用视频预测推理未观测状态，Rynn-GEU均分提升6.0分
- 控制闭环验证：World-Action扩展在RoboCasa365复合任务成功率51.5%，真机任务达46.7%
工程与产业观察
RynnUnify验证了单一模型同时承载高阶认知与底层控制的可行性，其交错视频-语言推理与RL增强机制为具身大模型从“看懂”迈向“做对”提供工程路径。46.7%的真机成功率表明统一架构已具备初步实操价值，对降低多模块系统集成复杂度具有参考意义。
英文摘要 (Abstract)
We introduce RynnUnify, a 16B unified embodied model that combines strong embodied understanding with long-horizon visual planning through interleaved video--language reasoning. With capabilities spanning spatial reasoning, physical prediction, and grounding, RynnUnify connects an understanding of the physical world with visual anticipation of how it evolves across successive subtasks. Building on these capabilities, we further develop an agentic extension through reinforcement learning to selectively use video predictions for reasoning about unobserved states and interaction outcomes. We also develop a world-action extension of the full RynnUnify model, translating its understanding and prediction capabilities into robot control. To evaluate these capabilities, we introduce Rynn-VP and Rynn-GEU for long-horizon visual planning and anticipatory reasoning. Experiments show that RynnUnify achieves the highest average embodied understanding score among the evaluated unified multimodal models and improves the Rynn-VP overall score by 9.75 points over Cosmos 3 Nano. Its agentic extension improves the Rynn-GEU average by 6.0 points over direct reasoning, while the world-action extension achieves 51.5% success on 20 selected composite RoboCasa365 tasks and 46.7% average success across three real-robot tasks. Together, these results establish RynnUnify as a shared foundation for embodied understanding, visual planning, and robot control, taking a step toward a unified embodied brain.
Hugging Face Daily Papers 热度：29 upvotes 作者：Minghao Zhu, Bohan Hou, Kun Xiang, Zixuan Wang, Yinzhou Tang, Yifei Yang, Zhikai Wang, Tianyi Zhang, Kehan Li, Haoyu Zhao, Siteng Huang, Leigang Qu, Sicong Leng, Dongchi Huang, Mingxiu Chen, Shijian L…
来源
本页事实对应的公开页面，以原文为准。