Reproduce “Wasserstein distance estimates for the distributions of numerical approximations to ergodic stochastic differential equations” 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

J.M. Sanz-Serna, Konstantinos C. Zygalakis

The paper bounds the 2-Wasserstein error between an exact SDE’s invariant distribution and a discretized chain by combining numerical contractivity with one-step strong error. Applied to underdamped Langevin, UBU uses one gradient per step and has scaling proportional to \((m^{1/2}\epsilon)^{-1/2}\kappa^{5/4}d^{1/4}\) under a corrected strongly Hessian-Lipschitz condition. Without it, UBU matches EE’s \(\epsilon^{-1}\kappa^{3/2}d^{1/2}\) scaling.

Summary by alphaXiv.

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