Hi-WM uses human interventions inside an action-conditioned world model with rollback and branching to generate dense corrective data, raising real-world success by 37.9 points on average across three manipulation tasks.
Precise and dexterous robotic manipulation via human- in-the-loop reinforcement learning
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.RO 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Hi-WM: Human-in-the-World-Model for Scalable Robot Post-Training
Hi-WM uses human interventions inside an action-conditioned world model with rollback and branching to generate dense corrective data, raising real-world success by 37.9 points on average across three manipulation tasks.