Reproduce “Active Exploration for Robotic Manipulation” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Tim Schneider, Boris Belousov, Georgia Chalvatzaki, Diego Romeres, Devesh K. Jha, Jan Peters
The authors present an information-seeking model-predictive-control method for balancing and pushing a ball toward a hidden target on a tilted table. It plans actions for task reward and expected information gain using an ensemble of models. In simulation, it solved Tilted Pushing where PETS, SAC, and MBPO failed within 300,000 steps; real-robot evidence is limited to this task.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.