Research map / Probabilistic and scientific modeling / Variational and generative sampling
Estimating Convergence of Markov chains with L-Lag Couplings
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Increase the coupling lag from L=1 to L=150 | L-lag coupling bounds for random-walk Metropolis-Hastings · Univariate Normal target | total variation and 1-Wasserstein upper-bound tightness | Helped |
| Use a larger coupling lag | L-lag coupling bounds · General Markov chains | upper-bound tightness | Helped |
| Use parallel tempering instead of single-site Gibbs | Ising model samplers at inverse temperature β = 0.46 · 32 × 32 Ising model | total variation upper-bound convergence | Helped |
| Use the Pólya-Gamma Gibbs sampler instead of the tested HMC samplers | Bayesian logistic regression samplers · German Credit | total variation upper bound | Helped |
| Increase the number of SMC particles | SMC particle approximation · SMC target-particle marginal | total variation upper bound | Helped |
| Estimate the 1-lag bound using only 1000 repeats in the bimodal example | L-lag coupling bound estimation · Bimodal target | reliability of the estimated total variation upper bound | Hurt |
Closest papers by meaning
- Multi-Marginal Couplings for Metropolis--Hastings (NeurIPS 2026 · similarity 0.93)
- Scalable Metropolis-Hastings for Exact Bayesian Inference with Large Datasets (ICML 2019 · similarity 0.88)
- Non-asymptotic Error Bounds in $\mathcal{W}_2$-Distance with Sqrt(d) Dimension Dependence and First Order Convergence for Langevin Monte Carlo beyond Log-Concavity (ICML 2025 · similarity 0.88)
- Sqrt(d) Dimension Dependence of Langevin Monte Carlo (ICLR 2022 · similarity 0.88)
- A Divergence Bound for Hybrids of MCMC and Variational Inference and an Application to Langevin Dynamics and SGVI (ICML 2017 · similarity 0.87)
- Parallelizing MCMC Across the Sequence Length (NeurIPS 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.