Annals of Applied Probability

Spectral gaps for a Metropolis–Hastings algorithm in infinite dimensions

Martin Hairer, Andrew M. Stuart, and Sebastian J. Vollmer

Full-text: Open access

Enhanced PDF (374 KB)

Abstract

We study the problem of sampling high and infinite dimensional target measures arising in applications such as conditioned diffusions and inverse problems. We focus on those that arise from approximating measures on Hilbert spaces defined via a density with respect to a Gaussian reference measure. We consider the Metropolis–Hastings algorithm that adds an accept–reject mechanism to a Markov chain proposal in order to make the chain reversible with respect to the target measure. We focus on cases where the proposal is either a Gaussian random walk (RWM) with covariance equal to that of the reference measure or an Ornstein–Uhlenbeck proposal (pCN) for which the reference measure is invariant.

Previous results in terms of scaling and diffusion limits suggested that the pCN has a convergence rate that is independent of the dimension while the RWM method has undesirable dimension-dependent behaviour. We confirm this claim by exhibiting a dimension-independent Wasserstein spectral gap for pCN algorithm for a large class of target measures. In our setting this Wasserstein spectral gap implies an $L^{2}$-spectral gap. We use both spectral gaps to show that the ergodic average satisfies a strong law of large numbers, the central limit theorem and nonasymptotic bounds on the mean square error, all dimension independent. In contrast we show that the spectral gap of the RWM algorithm applied to the reference measures degenerates as the dimension tends to infinity.

Article information

Source
Ann. Appl. Probab., Volume 24, Number 6 (2014), 2455-2490.

Dates
First available in Project Euclid: 26 August 2014

Permanent link to this document
https://projecteuclid.org/euclid.aoap/1409058037

Digital Object Identifier
doi:10.1214/13-AAP982

Mathematical Reviews number (MathSciNet)
MR3262508

Zentralblatt MATH identifier
1307.65002

Subjects
Primary: 65C40: Computational Markov chains 60B10: Convergence of probability measures 60J05: Discrete-time Markov processes on general state spaces 60J22: Computational methods in Markov chains [See also 65C40]

Keywords
Wasserstein spectral gaps $L^{2}$-spectral gaps Markov chain Monte Carlo in infinite dimensions weak Harris theorem random walk Metropolis

Citation

Hairer, Martin; Stuart, Andrew M.; Vollmer, Sebastian J. Spectral gaps for a Metropolis–Hastings algorithm in infinite dimensions. Ann. Appl. Probab. 24 (2014), no. 6, 2455--2490. doi:10.1214/13-AAP982. https://projecteuclid.org/euclid.aoap/1409058037


Export citation

References

  • Adler, R. J. (1990). An Introduction to Continuity, Extrema, and Related Topics for General Gaussian Processes. Institute of Mathematical Statistics Lecture Notes—Monograph Series 12. IMS, Hayward, CA.
    Mathematical Reviews (MathSciNet): MR1088478
  • Athreya, K. B. and Ney, P. (1978). A new approach to the limit theory of recurrent Markov chains. Trans. Amer. Math. Soc. 245 493–501.
    Mathematical Reviews (MathSciNet): MR511425
    Digital Object Identifier: doi:10.1090/S0002-9947-1978-0511425-0
  • Bakry, D. and Émery, M. (1985). Diffusions hypercontractives. In Séminaire de Probabilités, XIX, 1983/84. Lecture Notes in Math. 1123 177–206. Springer, Berlin.
    Mathematical Reviews (MathSciNet): MR889476
    Digital Object Identifier: doi:10.1007/BFb0075847
  • Beskos, A., Kalogeropoulos, K. and Pazos, E. (2013). Advanced MCMC methods for sampling on diffusion pathspace. Stochastic Process. Appl. 123 1415–1453.
    Mathematical Reviews (MathSciNet): MR3016228
    Digital Object Identifier: doi:10.1016/j.spa.2012.12.001
  • Beskos, A., Roberts, G. and Stuart, A. (2009). Optimal scalings for local Metropolis–Hastings chains on nonproduct targets in high dimensions. Ann. Appl. Probab. 19 863–898.
    Mathematical Reviews (MathSciNet): MR2537193
    Digital Object Identifier: doi:10.1214/08-AAP563
    Project Euclid: euclid.aoap/1245071013
  • Beskos, A., Roberts, G., Stuart, A. and Voss, J. (2008). MCMC methods for diffusion bridges. Stoch. Dyn. 8 319–350.
    Mathematical Reviews (MathSciNet): MR2444507
    Digital Object Identifier: doi:10.1142/S0219493708002378
  • Beskos, A., Pinski, F., Sanz-Serna, J. M. and Stuart, A. M. (2011). Hybrid Monte-Carlo on Hilbert spaces. Stochastic Process. Appl. 121 2201–2230.
    Mathematical Reviews (MathSciNet): MR2822774
    Digital Object Identifier: doi:10.1016/j.spa.2011.06.003
  • Bogachev, V. I. (1998). Gaussian Measures. Mathematical Surveys and Monographs 62. Amer. Math. Soc., Providence, RI.
    Mathematical Reviews (MathSciNet): MR1642391
  • Bogachev, V. I. (2007). Measure Theory. Vol. I, II. Springer, Berlin.
    Mathematical Reviews (MathSciNet): MR2267655
  • Chan, K. S. and Geyer, C. J. (1994). Discussion: Markov chains for exploring posterior distributions. Ann. Statist. 22 1747–1758.
    Mathematical Reviews (MathSciNet): MR1329166
    Digital Object Identifier: doi:10.1214/aos/1176325750
    Project Euclid: euclid.aos/1176325750
  • Cheeger, J. (1970). A lower bound for the smallest eigenvalue of the Laplacian. In Problems in Analysis (Papers Dedicated to Salomon Bochner, 1969) 195–199. Princeton Univ. Press, Princeton, NJ.
    Mathematical Reviews (MathSciNet): MR402831
  • Cotter, S. L., Roberts, G. O., Stuart, A. M. and White, D. (2013). MCMC methods for functions: Modifying old algorithms to make them faster. Statist. Sci. 28 424–446.
    Mathematical Reviews (MathSciNet): MR3135540
    Digital Object Identifier: doi:10.1214/13-STS421
    Project Euclid: euclid.ss/1377696944
  • Cuny, C. and Lin, M. (2009). Pointwise ergodic theorems with rate and application to the CLT for Markov chains. Ann. Inst. Henri Poincaré Probab. Stat. 45 710–733.
    Mathematical Reviews (MathSciNet): MR2548500
    Digital Object Identifier: doi:10.1214/08-AIHP180
    Project Euclid: euclid.aihp/1249391381
  • Dashti, M., Harris, S. and Stuart, A. (2012). Besov priors for Bayesian inverse problems. Inverse Probl. Imaging 6 183–200.
    Mathematical Reviews (MathSciNet): MR2942737
    Digital Object Identifier: doi:10.3934/ipi.2012.6.183
  • Dashti, M. and Stuart, A. M. (2011). Uncertainty quantification and weak approximation of an elliptic inverse problem. SIAM J. Numer. Anal. 49 2524–2542.
    Mathematical Reviews (MathSciNet): MR2873245
    Digital Object Identifier: doi:10.1137/100814664
  • Da Prato, G. and Zabczyk, J. (1992). Stochastic Equations in Infinite Dimensions. Encyclopedia of Mathematics and Its Applications 44. Cambridge Univ. Press, Cambridge.
    Mathematical Reviews (MathSciNet): MR1207136
  • Diaconis, P. and Stroock, D. (1991). Geometric bounds for eigenvalues of Markov chains. Ann. Appl. Probab. 1 36–61.
    Mathematical Reviews (MathSciNet): MR1097463
    Digital Object Identifier: doi:10.1214/aoap/1177005980
    Project Euclid: euclid.aoap/1177005980
  • Eberle, A. (2014). Error bounds for Metropolis–Hastings algorithms applied to perturbations of Gaussian measures in high dimensions. Ann. Appl. Probab. 24 337–377.
    Mathematical Reviews (MathSciNet): MR3161650
    Digital Object Identifier: doi:10.1214/13-AAP926
    Project Euclid: euclid.aoap/1389278728
  • Frigessi, A., di Stefano, P., Hwang, C.-R. and Sheu, S. J. (1993). Convergence rates of the Gibbs sampler, the Metropolis algorithm and other single-site updating dynamics. J. Roy. Statist. Soc. Ser. B 55 205–219.
    Mathematical Reviews (MathSciNet): MR1210432
  • Geyer, C. J. (1992). Practical Markov chain Monte Carlo. Statist. Sci. 7 473–483.
  • Geyer, C. J. and Thompson, E. A. (1995). Annealing Markov chain Monte Carlo with applications to ancestral inference. J. Amer. Statist. Assoc. 90 909–920.
  • Hairer, M. (2010). An introduction to stochastic PDEs. Lecture notes, University of Warwick.
  • Hairer, M. and Majda, A. J. (2010). A simple framework to justify linear response theory. Nonlinearity 23 909–922.
    Mathematical Reviews (MathSciNet): MR2602020
    Digital Object Identifier: doi:10.1088/0951-7715/23/4/008
  • Hairer, M., Mattingly, J. C. and Scheutzow, M. (2011). Asymptotic coupling and a general form of Harris’ Theorem with applications to stochastic delay equations. Probab. Theory Related Fields 149 223–259.
    Mathematical Reviews (MathSciNet): MR2773030
    Digital Object Identifier: doi:10.1007/s00440-009-0250-6
  • Hairer, M., Stuart, A. M. and Voss, J. (2007). Analysis of SPDEs arising in path sampling. II. The nonlinear case. Ann. Appl. Probab. 17 1657–1706.
    Mathematical Reviews (MathSciNet): MR2358638
    Digital Object Identifier: doi:10.1214/07-AAP441
    Project Euclid: euclid.aoap/1191419180
  • Hastings, W. K. (1970). Monte-Carlo sampling methods using Markov chains and their applications. Biometrika 57 97.
  • Hjort, N. L., Holmes, C., Müller, P. and Walker, S. G., eds. (2010). Bayesian Nonparametrics. Cambridge Series in Statistical and Probabilistic Mathematics 28. Cambridge Univ. Press, Cambridge.
    Mathematical Reviews (MathSciNet): MR2722987
  • Joulin, A. and Ollivier, Y. (2010). Curvature, concentration and error estimates for Markov chain Monte Carlo. Ann. Probab. 38 2418–2442.
    Mathematical Reviews (MathSciNet): MR2683634
    Digital Object Identifier: doi:10.1214/10-AOP541
    Project Euclid: euclid.aop/1285334210
  • Kipnis, C. and Varadhan, S. R. S. (1986). Central limit theorem for additive functionals of reversible Markov processes and applications to simple exclusions. Comm. Math. Phys. 104 1–19.
    Mathematical Reviews (MathSciNet): MR834478
    Digital Object Identifier: doi:10.1007/BF01210789
    Project Euclid: euclid.cmp/1104114929
  • Komorowski, T. and Walczuk, A. (2012). Central limit theorem for Markov processes with spectral gap in the Wasserstein metric. Stochastic Process. Appl. 122 2155–2184.
    Mathematical Reviews (MathSciNet): MR2921976
    Digital Object Identifier: doi:10.1016/j.spa.2012.03.006
  • Lassas, M., Saksman, E. and Siltanen, S. (2009). Discretization-invariant Bayesian inversion and Besov space priors. Inverse Probl. Imaging 3 87–122.
    Mathematical Reviews (MathSciNet): MR2558305
    Digital Object Identifier: doi:10.3934/ipi.2009.3.87
  • Łatuszyński, K. and Niemiro, W. (2011). Rigorous confidence bounds for MCMC under a geometric drift condition. J. Complexity 27 23–38.
    Mathematical Reviews (MathSciNet): MR2745298
    Digital Object Identifier: doi:10.1016/j.jco.2010.07.003
  • Łatuszyński, K. and Roberts, G. O. (2013). CLTs and asymptotic variance of time-sampled Markov chains. Methodol. Comput. Appl. Probab. 15 237–247.
    Mathematical Reviews (MathSciNet): MR3030220
    Digital Object Identifier: doi:10.1007/s11009-011-9237-8
  • Lawler, G. F. and Sokal, A. D. (1988). Bounds on the $L^{2}$ spectrum for Markov chains and Markov processes: A generalization of Cheeger’s inequality. Trans. Amer. Math. Soc. 309 557–580.
    Mathematical Reviews (MathSciNet): MR930082
  • Lee, P. M. (2004). Bayesian Statistics: An Introduction, 3rd ed. Arnold, London.
    Mathematical Reviews (MathSciNet): MR2069264
  • Liu, J. S. (2008). Monte Carlo Strategies in Scientific Computing. Springer, New York.
    Mathematical Reviews (MathSciNet): MR2401592
  • Lovász, L. and Simonovits, M. (1993). Random walks in a convex body and an improved volume algorithm. Random Structures Algorithms 4 359–412.
    Mathematical Reviews (MathSciNet): MR1238906
  • Mattingly, J. C., Pillai, N. S. and Stuart, A. M. (2012). Diffusion limits of the random walk Metropolis algorithm in high dimensions. Ann. Appl. Probab. 22 881–930.
    Mathematical Reviews (MathSciNet): MR2977981
    Digital Object Identifier: doi:10.1214/10-AAP754
    Project Euclid: euclid.aoap/1337347534
  • Metropolis, N., Rosenbluth, A. W., Rosenbluth, M. N., Teller, A. H., Teller, E. et al. (1953). Equation of state calculations by fast computing machines. J. Chem. Phys. 21 1087.
  • Meyn, S. and Tweedie, R. L. (2009). Markov Chains and Stochastic Stability, 2nd ed. Cambridge Univ. Press, Cambridge.
    Mathematical Reviews (MathSciNet): MR2509253
  • Nummelin, E. (1978). A splitting technique for Harris recurrent Markov chains. Probab. Theory Related Fields 43 309–318.
    Mathematical Reviews (MathSciNet): MR501353
  • Pillai, N. S., Stuart, A. M. and Thiéry, A. H. (2011). Optimal proposal design for random walk type Metropolis algorithms with Gaussian random field priors. ArXiv E-prints.
  • Robert, C. P. and Casella, G. (2004). Monte Carlo Statistical Methods, 2nd ed. Springer, New York.
    Mathematical Reviews (MathSciNet): MR2080278
  • Roberts, G. O. and Tweedie, R. L. (1996). Geometric convergence and central limit theorems for multidimensional Hastings and Metropolis algorithms. Biometrika 83 95–110.
    Mathematical Reviews (MathSciNet): MR1399158
    Digital Object Identifier: doi:10.1093/biomet/83.1.95
  • Röckner, M. and Wang, F.-Y. (2001). Weak Poincaré inequalities and $L^{2}$-convergence rates of Markov semigroups. J. Funct. Anal. 185 564–603.
    Mathematical Reviews (MathSciNet): MR1856277
    Digital Object Identifier: doi:10.1006/jfan.2001.3776
  • Rudolf, D. (2012). Explicit error bounds for Markov chain Monte Carlo. Dissertationes Math. (Rozprawy Mat.) 485 1–93.
    Mathematical Reviews (MathSciNet): MR2977521
    Digital Object Identifier: doi:10.4064/dm485-0-1
  • Schwab, C. and Stuart, A. M. (2012). Sparse deterministic approximation of Bayesian inverse problems. Inverse Problems 28 045003, 32.
    Mathematical Reviews (MathSciNet): MR2903278
  • Sinclair, A. and Jerrum, M. (1989). Approximate counting, uniform generation and rapidly mixing Markov chains. Inform. and Comput. 82 93–133.
    Mathematical Reviews (MathSciNet): MR1003059
    Digital Object Identifier: doi:10.1016/0890-5401(89)90067-9
  • Stuart, A. M. (2010). Inverse problems: A Bayesian perspective. Acta Numer. 19 451–559.
    Mathematical Reviews (MathSciNet): MR2652785
    Digital Object Identifier: doi:10.1017/S0962492910000061
  • Tierney, L. (1998). A note on Metropolis–Hastings kernels for general state spaces. Ann. Appl. Probab. 8 1–9.
    Mathematical Reviews (MathSciNet): MR1620401
    Digital Object Identifier: doi:10.1214/aoap/1027961031
    Project Euclid: euclid.aoap/1027961031
  • Vollmer, S. J. (2013). Dimension-independent MCMC sampling for inverse problems with non-Gaussian priors. Available at arXiv:1302.2213.
    arXiv: 1302.2213
    Mathematical Reviews (MathSciNet): MR3141858
  • Wang, F.-Y. (2003). Functional inequalities for the decay of sub-Markov semigroups. Potential Anal. 18 1–23.
    Mathematical Reviews (MathSciNet): MR1953493
    Digital Object Identifier: doi:10.1023/A:1020535718522

The Institute of Mathematical Statistics

Annals of Applied Probability

New content alerts

Email RSS ToC RSS Article
  • You have access to this content.
  • You have partial access to this content.
  • You do not have access to this content.
Turn Off MathJax

What is MathJax?

More like this

+ See more