Changhao Wang
@ChanghaoWang78
Followers
48
Following
65
Media
0
Statuses
9
Meta FAIR Robotics, Ph.D. UC Berkeley
Menlo Park
Joined August 2021
Human videos are great for robot learning, but they critically lack the sense of touch needed for complex manipulation tasks! Introducing OSMO, an open-source tactile glove that captures rich contact signals during human demonstrations for direct, successful transfer to robots.
5
31
166
Massive credit to @irmakkguzey for designing and implementing the entire pipeline. Seeing robots learn dexterous skills purely from Aria-glass footage still feels a bit surreal 😁
Dexterous manipulation by directly observing humans - a dream in AI for decades - is hard due to visual and embodiment gaps. With simple yet powerful hardware - Aria 2 glasses 👓 - and our new work AINA 🪞, we are now one significant step closer to achieving this dream.
0
0
4
We believe dexterity naturally comes from humans. SPIDER offers an efficient way to convert large-scale human video data into dynamically feasible robot data. Check out our website: https://t.co/ZzcnUe3RcQ Code will be released soon:
github.com
A general physic-based retargeting framework. . Contribute to jc-bao/spider development by creating an account on GitHub.
🕸️ Introducing SPIDER — Scalable Physics-Informed Dexterous Retargeting! A dynamically feasible, cross-embodiment retargeting framework for BOTH humanoids 🤖 and dexterous hands ✋. From human motion → sim → real robots, at scale. 🔗 Website: https://t.co/ieZfG2Q4L0 🧵 1/n
0
0
5
Just open-sourced Geometric Retargeting (GeoRT) — the kinematic retargeting module behind DexterityGen. Includes tools for importing custom hands. Give it a try: https://t.co/MxSuitRaDM A software by @berkeley_ai and @AIatMeta. More coming soon.
1
13
76
We consider teleoperation as a prompting problem, where the human action serves as a prompt to our foundation model, Dexterity Gen. Our model then generates the optimal action that ensures both stability and preservation of the human intent. Checkout our paper for more details.
We introduce Dexterity Gen (DexGen), a foundation controller that enables unprecedented dexterous manipulation capabilities. For the first time, it allows human teleoperation of tasks such as using a pen, screwdriver, and syringe. Developed by @berkeley_AI and @MetaAI. A Thread.
0
3
11