Research organization
Berkeley AI Research
Robot learning, control, reinforcement learning
Official research siteDirect answer
What does Berkeley AI Research research?
Robot learning, control, reinforcement learning
Physical AI work
Berkeley AI Research contributes to the Physical AI ecosystem through robot learning, control, reinforcement learning. This profile connects its public research projects with the roadmap topics needed to understand them.
Notable projects
- OpenVLA
- BridgeData
- RoboNet
Related paper explainers
OpenVLA: An Open-Source Vision-Language-Action Model
OpenVLA is an open 7B-parameter VLA trained on the Open X-Embodiment dataset.
VLAOpen Source
Read explanation DROID: Distributed Robot Interaction Dataset
DROID provides diverse in-the-wild robot manipulation demonstrations across many scenes, tasks, and collectors.
DatasetsManipulationRobot Learning
Read explanation