Reproduce “Back to the Future: Predicting Long-Horizon Dynamics with Retrospective Correction for Disturbance-Aware Retrieval” 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
Jiwoong Choi, Haoyu Li, Yiwei Lyu
DARE predicts geometry-aware object interactions and uses later observations to revise an earlier predicted state, rerunning dynamics before planning further contact. Its planner balances retrieval progress against predicted bystander disturbance. In simulated clean-observation cabinet scenes with 10-object hard layouts, DARE achieved 1.00 success versus 0.20 for direct retrieval, with lower reported translation, tilt, and yaw disturbance; online correction was off.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.