TY  - RPRT
TI  - Uncertainty-Aware Robotic World Model Makes Offline Model-Based Reinforcement Learning Work on Real Robots
AU  - Chenhao Li
AU  - Andreas Krause
AU  - Marco Hutter
PY  - 2026
UR  - https://arxiv.org/abs/2504.16680
ID  - 2504.16680
ER  -