Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
173 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
46 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Scalable couplings for the random walk Metropolis algorithm (2211.12585v3)

Published 22 Nov 2022 in stat.CO, math.PR, and stat.ME

Abstract: There has been a recent surge of interest in coupling methods for Markov chain Monte Carlo algorithms: they facilitate convergence quantification and unbiased estimation, while exploiting embarrassingly parallel computing capabilities. Motivated by these, we consider the design and analysis of couplings of the random walk Metropolis algorithm which scale well with the dimension of the target measure. Methodologically, we introduce a low-rank modification of the synchronous coupling that is provably optimally contractive in standard high-dimensional asymptotic regimes. We expose a shortcoming of the reflection coupling, the state of the art at the time of writing, and we propose a modification which mitigates the issue. Our analysis bridges the gap to the optimal scaling literature and builds a framework of asymptotic optimality which may be of independent interest. We illustrate the applicability of our proposed couplings, and the potential for extending our ideas, with various numerical experiments.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (56)
  1. The pseudo-marginal approach for efficient Monte Carlo computations. The Annals of Statistics, 37(2):697 – 725, 2009.
  2. Explicit convergence bounds for Metropolis Markov chains: isoperimetry, spectral gaps and profiles. The Annals of Applied Probability (to appear), 2024.
  3. Optimal tuning of the hybrid Monte Carlo algorithm. Bernoulli, 19(5A):1501 – 1534, 2013.
  4. Asymptotic analysis of the random walk Metropolis algorithm on ridged densities. The Annals of Applied Probability, 28(5):2966 – 3001, 2018.
  5. Bounding Wasserstein Distance with Couplings. Journal of the American Statistical Association, 0(0):1–12, 2024.
  6. Estimating convergence of Markov chains with L-lag couplings. In Neural Information Processing Systems (NeurIPS), 2019.
  7. Coupling-based convergence assessment of some Gibbs samplers for high-dimensional Bayesian regression with shrinkage priors. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 84(3):973–996, 2022.
  8. Coupling and convergence for Hamiltonian Monte Carlo. The Annals of Applied Probability, 30(3):1209 – 1250, 2020.
  9. Coupling Methods for Multidimensional Diffusion Processes. The Annals of Probability, 17(1):151 – 177, 1989.
  10. Leave Pima Indians Alone: Binary Regression as a Benchmark for Bayesian Computation. Statistical Science, 32(1):64 – 87, 2017.
  11. Scaling limits for the transient phase of local Metropolis–Hastings algorithms. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 67(2):253–268, 2005.
  12. Stephen Connor. Coupling: Cutoffs, CFTP and Tameness. PhD thesis, University of Warwick, 2007.
  13. The coupled rejection sampler. arXiv preprint arXiv:2201.09585, 2022.
  14. Solving the Poisson equation using coupled Markov chains. arXiv preprint arXiv:2206.05691, 2023.
  15. Hybrid Monte Carlo. Physics Letters B, 195(2):216–222, 1987.
  16. Andreas Eberle. Error bounds for Metropolis–Hastings algorithms applied to perturbations of Gaussian measures in high dimensions. The Annals of Applied Probability, 24(1):337–377, 2014.
  17. Matthias Gelbrich. On a formula for the L2 Wasserstein metric between measures on Euclidean and Hilbert spaces. Mathematische Nachrichten, 147(1):185–203, 1990.
  18. A weakly informative default prior distribution for logistic and other regression models. The Annals of Applied Statistics, 2(4):1360 – 1383, 2008.
  19. J Heng and P E Jacob. Unbiased Hamiltonian Monte Carlo with couplings. Biometrika, 106(2):287–302, 2019.
  20. Better together? Statistical learning in models made of modules. arXiv preprint arxiv:1708.08719, 2017.
  21. Unbiased Markov chain Monte Carlo methods with couplings. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 82(3):543–600, 2020.
  22. Diffusion limit for the random walk Metropolis algorithm out of stationarity. Annales de l’Institut Henri Poincaré, Probabilités et Statistiques, 55(3):1599 – 1648, 2019.
  23. Jun S. Liu. Monte Carlo Strategies in Scientific Computing. Springer, 2001.
  24. The Barker Proposal: Combining Robustness and Efficiency in Gradient-Based MCMC. Journal of the Royal Statistical Society Series B: Statistical Methodology, 84(2):496–523, 2022.
  25. M Ludkin and C Sherlock. Hug and hop: a discrete-time, nonreversible Markov chain Monte Carlo algorithm. Biometrika, 110(2):301–318, 07 2022.
  26. Unbiased Markov chain Monte Carlo for intractable target distributions. Electronic Journal of Statistics, 14(2):2842 – 2891, 2020.
  27. Radford M. Neal. MCMC Using Hamiltonian Dynamics. In Handbook of Markov Chain Monte Carlo, chapter 5. CRC Press, 2011.
  28. John O’Leary. Couplings of the Random-Walk Metropolis algorithm. arXiv preprint arXiv:2102.01790, 2021.
  29. Maximal Couplings of the Metropolis-Hastings Algorithm. In International Conference on Artificial Intelligence and Statistics (AISTATS), 2021.
  30. Bounds on Wasserstein distances between continuous distributions using independent samples. arXiv preprint arXiv::2203.11627, 2022.
  31. Optimal Scaling of Discrete Approximations to Langevin Diffusions. Journal of the Royal Statistical Society Series B: Statistical Methodology, 60(1):255–268, 01 1998.
  32. Optimal scaling for various Metropolis-Hastings algorithms. Statistical Science, 16(4):351 – 367, 2001.
  33. Weak convergence and optimal scaling of random walk Metropolis algorithms. The Annals of Applied Probability, 7(1):110–120, 1997.
  34. Gaussian Markov Random Fields: Theory and Applications. Chapman and Hall/CRC, 2005.
  35. Chris Sherlock. Optimal scaling of the random walk Metropolis: general criteria for the 0.234 acceptance rule. Journal of Applied Probability, 50(1):1 – 15, 2013.
  36. Optimal scaling of the random walk Metropolis on elliptically symmetric unimodal targets. Bernoulli, 15(3):774 – 798, 2009.
  37. Cédric Villani. Optimal Transport: Old and New. Springer, 2009.
  38. Metropolis-Hastings transition kernel couplings. Annales de l’Institut Henri Poincaré, Probabilités et Statistiques (to appear), 2023.
  39. Unbiased Multilevel Monte Carlo Methods for Intractable Distributions: MLMC Meets MCMC. Journal of Machine Learning Research, 24(249):1–40, 2023.
  40. On the asymptotic behavior of the expectation of the maximum of i.i.d. random variables. Operations Research Letters, 49(5):785–786, 2021.
  41. Peter J. Downey. Distribution-free bounds on the expectation of the maximum with scheduling applications. Operations Research Letters, 9(3):189–201, 1990.
  42. The jackknife estimate of variance. The Annals of Statistics, 9(3):586–596, 1981.
  43. mcmcse: Monte Carlo Standard Errors for MCMC. Riverside, CA, and Kanpur, India, 2021. R package version 1.5-0.
  44. mvtnorm: Multivariate Normal and t Distributions, 2021. R package version 1.1-3.
  45. Eli. Grushka. Characterization of exponentially modified gaussian peaks in chromatography. Analytical Chemistry, 44(11):1733–1738, 1972.
  46. Olav Kallenberg. Foundations of Modern Probability. Springer, 2021.
  47. Donald Bruce Owen. A table of normal integrals. Communications in Statistics - Simulation and Computation, 9(4):389–419, 1980.
  48. Coda: Convergence diagnosis and output analysis for mcmc. R News, 6(1):7–11, 2006.
  49. R Core Team. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria, 2020.
  50. Limit laws of the empirical Wasserstein distance: Gaussian distributions. Journal of Multivariate Analysis, 151:90–109, 2016.
  51. Ken-iti Sato. Lévy Processes and Infinitely Divisible Distributions. Cambridge Studies in Advanced Mathematics. Cambridge University Press, 1999.
  52. Solving differential equations in R: Package deSolve. Journal of Statistical Software, 33(9):1–25, 2010.
  53. William F. Trench. Asymptotic distribution of the spectra of a class of generalized Kac–Murdock–Szegö matrices. Linear Algebra and its Applications, 294(1):181–192, 1999.
  54. D Vats and J M Flegal. Lugsail lag windows for estimating time-average covariance matrices. Biometrika, 109(3):735–750, 10 2021.
  55. Cédric Villani. Topics in Optimal Transportation. Graduate studies in mathematics. American Mathematical Society, 2003.
  56. Bernt Øksendal. Stochastic Differential Equations. Springer-Verlag, 5 edition, 1998.
Citations (1)
View on Semantic Scholar

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now