Reproduce “Underdamped Anchored Langevin Dynamics” 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
Qi Feng, Hoang M. Nguyen, Lingjiong Zhu
UALMC extends anchored Langevin sampling to nonsmooth energies, using a smooth reference gradient while scaling force and position speed to preserve the original target in continuous time. Under regularity and small-mismatch conditions, its iteration bound scales as square root of dimension over accuracy. At dimension 10 and step size 0.4, its measured bias was 0.038 versus ALMC’s 0.280; finite-step bias remains.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.