Reproduce “Grounded-Dreamer: Robot Learning from World Model Synthetic Data with Physical Grounding” 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

Feng Jiang, Shu Chen, Yuanfei Wang, Yuan Li, Zhenhao Shen, Jiaqi Liang, Yang Chen, Lei Wei and 2 more

Grounded-Dreamer creates robot training trajectories from real reference images. It combines a physics simulator with a video world model, infers actions from generated videos using an inverse dynamics model, and replays them in simulation, retaining only trajectories passing task checks. On Press Button, π0.5 trained on its data achieved 75% success, matching the real-only policy in the tested comparison.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.