Event

OpenAI trains robotic hand to solve Rubik's Cube using Automatic Domain Randomization

Oct 15, 2019 · Research Publication · OpenAI · Foundation Models

  • OpenAI published robotic-hand work showing neural networks trained in simulation solving a Rubik's Cube on real hardware.
  • The project introduced Automatic Domain Randomization to improve sim-to-real transfer for physical manipulation.
  • OpenAI framed the work as a step toward general-purpose robots through software advances on existing robotic-hand hardware.

OpenAI develops foundation models, AI products and research systems, with a renewed robotics program focused on general-purpose robotics, robotic data acquisition, perception, simulation and model evaluation in physical environments. Its robotics relevance is model-layer and infrastructure-led rather than OEM-led.

Company
OpenAI
Segment
Foundation Models
Event type
Research Publication
Geography
San Francisco · United States