Video models serve as a good pretrained backbone for robot policies. Paper: htt...

Chelsea Finn (@chelseabfinn)

2026-01-24 | ❤️ 506 | 🔁 48 | 💬 8

---

Video models serve as a good pretrained backbone for robot policies.

Paper: https://arxiv.org/abs/2601.16163 Code: https://github.com/nvlabs/cosmos-policy

🔗 원본 링크

• https://arxiv.org/abs/2601.16163
• https://github.com/nvlabs/cosmos-policy

인용된 트윗

@moo_jin_kim: We release Cosmos Policy 💫: a state-of-the-art robot policy built on a video diffusion model backbone.

• policy + world model + value function — in 1 model
• no architectural changes to the base vide…

---

🔗 Related

• what-if-we-could-train-ai-robots-in-a-perfect-physics — 주제: AI-ML, Dev-Tools, Robotics, GenAI
• do-we-really-need-an-external-world-model-standard — 주제: AI-ML, Robotics, GenAI
• 16-ego-centric-world-models-we-introduce-egowm-a-video — 주제: AI-ML, Robotics, GenAI
• you-could-direct-a-video-like-a-real-3d-world-researchers — 주제: AI-ML, Dev-Tools, GenAI
• exciting-new-work-on-detailed-pixel-level-dense-3d-visual — 주제: AI-ML, Robotics, GenAI

인용 트윗

Moo Jin Kim (@moo_jin_kim)

We release Cosmos Policy 💫: a state-of-the-art robot policy built on a video diffusion model backbone.

• policy + world model + value function — in 1 model
• no architectural changes to the base video model
• SOTA in LIBERO (98.5%), RoboCasa (67.1%), & ALOHA tasks (93.6%)

🧵👇 https://t.co/cz9L3ziJ6x

원본 트윗

🎬 영상

Tags

Robotics GenAI AI-ML Dev-Tools