Introducing ZeroDex: Zero-Shot Long-Horizon Dexterous Manipulation via Multi-View 3D-Grounded VLM Reasoning
VLMs know what to do. ZeroDex grounds that reasoning in multi-view 3D, enabling robots to perform dexterous actions.
No task-specific demos. No policy fine-tuning.
Introducing ZeroDex: Zero-Shot Long-Horizon Dexterous Manipulation via Multi-View 3D-Grounded VLM Reasoning
VLMs know what to do. ZeroDex grounds that reasoning in multi-view 3D, enabling robots to perform dexterous actions.
No task-specific demos. No policy fine-tuning.
For long-horizon tasks, ZeroDex closes the loop with task verification and retry.
We demonstrate zero-shot dexterous manipulation across unseen objects, tool-use tasks, and long-horizon real-robot scenarios.
๐ Project: https://t.co/UXIkp5QHuA
๐ Paper: https://t.co/C36rTEcAxf
We introduce ๐๏ธ๐ ๐๐๐ฑ๐ญ๐๐ซ๐จ๐ฎ๐ฌ ๐๐จ๐ซ๐ฅ๐ ๐๐จ๐๐๐ฅ๐ฌ (๐๐๐) โ a scene-action-conditioned video diffusion model that simulates human manipulation in static 3D scenes from egocentric hand motions.
๐ Paper: https://t.co/u2IldpfYTD
๐ Project Page: https://t.co/YSE3rOmV1s
The full version of the ParaHome dataset is now available! ๐
Check out our webpage: https://t.co/qI2HWGQd2q
- 486 min, 207 sequences, 38 subjects
- 3D parameterized 22 objects
- Text annot. for every action
- Full 3D mocap for body and hand movements
- Sequential HOI scenarios