Reproduce “SMART: Zero-Shot Sim-to-Real Articulated Object Manipulation via Large-Scale Synthetic Pretraining” 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
Jicong Ao, Shuhan Jiang, Yuling Zhong, Yanwen Liu, Yuhan Gao, Jiangyuan Zhao, Yang Zhang, Shiqiang Zhu and 2 more
SMART-Sim generates articulated-object demonstrations in simulation by annotating functional parts, selecting contact points, and planning motions constrained by object joints. Its data trains policies across multiple robot setups. In real-robot evaluations, SMART-VLA achieved an average success rate of 76.7% across 12 task/platform settings, without real-world demonstrations. However, policies were post-trained on 500 task-specific simulated demonstrations before hardware evaluation.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.