Dexterity from Supervised Play

Key points

• How can robots autonomously collect the contact-rich experience needed to learn dexterous manipulation?1
• To accelerate inference, we also introduce a method for distilling a closed-loop, goal-conditioned policy through self-play within the learned world model.1
• We evaluate this framework on underactuated pushing and any-to-any in-hand pose reorientation with a multifingered hand, in simulation and real world.1
• These results demonstrate that supervised play can autonomously collect interaction data, guide model-based planning, and enable policy learning without any task-specific demonstrations or sim-to-real transfer.1

Sources (1)

• [1]Dexterity from Supervised Play
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 15, 05:55 PM

  We present DSP: Dexterity from supervised play, a framework in which a robot learns a play policy from minutes of human play through behavior cloning, then deploys it autonomously to collect hours of interaction data to train a world model.

  How can robots autonomously collect the contact-rich experience needed to learn dexterous manipulation?

Extractive summary: sentences quoted from the sources.