Reproduce “Stochastic Port-Hamiltonian Gaussian Processes for Learning from Noisy Dynamics on Lie Groups” 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
Thribhuvan Rapolu, Luke McLennan, Chandrajit Bajaj
LieSPHGP learns structured stochastic dynamics from noisy rigid-body trajectories, separating energy, dissipation, control inputs, physical forcing, and observation noise while preserving rotation geometry. On 10 held-out windy-pendulum trajectories, its 10-second rotation error is 0.291 rad, compared with 0.471 rad for a matched neural variant. In simulated swing-up, success is 100% at normal gains and 78% at lower gains, versus 84% for the true model.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.