Reproduce “Dimension-free regimes for unadjusted Langevin in ridge-separable models: A subspace robust Wasserstein analysis” 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
Yuansi Chen, Francesco Pedrotti, Peter A. Whalley
The authors analyze how Unadjusted Langevin (ULA) estimates posterior means under dimension-dependent error. They introduce subspace robust Wasserstein distance (SRW), which controls mean error and low-dimensional projections rather than requiring the full distribution to match. For ridge-separable models, iteration bounds are dimension-independent when data counts are small or sufficiently large relative to dimension; they fail in general near the proportional regime.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.