AI TrackerFrontier desk · AEST
Back to news
normalaitech

Dexterity from Supervised Play

arXiv cs.AI · 16 Oct 2026, 03:55 AEST

A new framework called DSP allows robots to learn dexterous manipulation through supervised play. This method significantly improves success rates while requiring much less interaction data compared to traditional reinforcement learning.

Key facts

Read aloud

A new framework called DSP allows robots to learn dexterous manipulation through supervised play. This method significantly improves success rates while requiring much less interaction data compared to traditional reinforcement learning. 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. In real world experiment, our method outperforms hindsight behavior cloning by sixty nine point six percent in success rate and achieves comparable success to reinforcement learning with four hundred ninety times less interaction data. 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.

Sources in this cluster