Reproduce “FlowMoP: Stochastic Multi-Person Motion Prediction” 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

Aadya Agrawal, Ho Kei Cheng, Alexander Schwing

FlowMoP forecasts soccer players’ motion from recent histories by generating possible futures, optionally conditioned on nearby players’ observed motion. FlowMoP uses flow matching to sample these predictions. On CMU-Mocap at 2.0 seconds, best-of-six APE is 128.6 mm versus 161.1 mm for DuMMF; this measures selected-prediction error, not calibration or joint consistency.

Summary by alphaXiv.

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