Reproduce “Making Koopman Control Guarantees Hold on Robots” 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

Abulikemu Abuduweili, Changliu Liu

TIK matches learned latent dynamics, to first order, with fitted plant dynamics near a target, addressing the mismatch between self-fed predictions and deployment’s re-encoding of measured states. Across four simulated open-loop-unstable benchmarks, 40 of 40 TIK controllers and 0 of 40 Standard Koopman controllers were locally stable on the plant; this is simulation evidence, not a general guarantee.

Summary by alphaXiv.

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