REGRIND learns dexterous manipulation policies from a single human demonstration

The authors present REGRIND, a minimalist retargeting-guided reinforcement learning pipeline that learns dexterous manipulation policies from a single human demonstration. The method retargets human hand-object motion to robot references while preserving spatial and contact relationships, then trains a residual RL policy in simulation to track object-centric keypoints.

• REGRIND transfers the resulting policy zero-shot to hardware using careful system identification.
• Policies produce fluid, human-like behavior on two different multi-fingered hands across contact-rich tool-use tasks.
• The work includes operating scissors and turning a screwdriver, with systematic hardware experiments analyzing sim-to-real transfer factors.

The authors offer practical guidance for retargeting-based learning in contact-rich settings by identifying the key factors that govern successful sim-to-real transfer.