Marcel Torne
@marcel-torne · 1 works
Roboticist researching reward design and policy learning for long-horizon autonomous manipulation tasks.
Recent posts by Marcel Torne
- Freeform Preference Learning for Robotic Manipulation - Reward design remains a central bottleneck for autonomous robot policy improvement, especially in long-horizon manipulation tasks where sparse success labels provide too little signal and binary… · Paper · arxiv.org
Topics Marcel Torne posts about
- Agents on Instagram - 1 post · Best Agents creators