Hybrid Monte Carlo methods for sampling probability measures on submanifolds. (English) Zbl 1453.65027
Numer. Math. 143, No. 2, 379-421 (2019); correction ibid. 144, No. 2, 447-449 (2020).
Summary: Probability measures supported on submanifolds can be sampled by adding an extra momentum variable to the state of the system, and discretizing the associated Hamiltonian dynamics with some stochastic perturbation in the extra variable. In order to avoid biases in the invariant probability measures sampled by discretizations of these stochastically perturbed Hamiltonian dynamics, a Metropolis rejection procedure can be considered. The so-obtained scheme belongs to the class of generalized Hybrid Monte Carlo (GHMC) algorithms. We show here how to generalize to GHMC a procedure suggested by Goodman, Holmes-Cerfon and Zappa for Metropolis random walks on submanifolds, where a reverse projection check is performed to enforce the reversibility of the algorithm for any timesteps and hence avoid biases in the invariant measure. We also provide a full mathematical analysis of such procedures, as well as numerical experiments demonstrating the importance of the reverse projection check on simple toy examples.
MSC:
| 65C40 | Numerical analysis or methods applied to Markov chains |
| 65P10 | Numerical methods for Hamiltonian systems including symplectic integrators |
| 70H45 | Constrained dynamics, Dirac’s theory of constraints |
References:
| [1] | Abraham, R.; Marsden, Je, Foundations of Mechanics (1978), San Francisco: Benjamin/Cummings Publishing Co. Inc., San Francisco · Zbl 0393.70001 |
| [2] | Afshar, H.M., Domke, J.: Reflection, refraction, and Hamiltonian Monte Carlo. In: Advances in Neural Information Processing Systems, pp. 3007-3015 (2015) |
| [3] | Andersen, Hc, Rattle: a “velocity” version of the Shake algorithm for molecular dynamics calculations, J. Comput. Phys., 52, 24-34 (1983) · Zbl 0513.65052 · doi:10.1016/0021-9991(83)90014-1 |
| [4] | Arnol’D, Vi, Mathematical Methods of Classical Mechanics (1989), Berlin: Springer, Berlin · Zbl 0692.70003 · doi:10.1007/978-1-4757-2063-1 |
| [5] | Bou-Rabee, N.; Vanden-Eijnden, E., Pathwise accuracy and ergodicity of metropolized integrators for SDEs, Commun. Pure Appl. Math., 63, 5, 655-696 (2009) · Zbl 1214.60031 |
| [6] | Breiding, P., Marigliano, O.: Sampling from the uniform distribution on an algebraic manifold. arXiv preprint arXiv:1810.06271 (2018) · Zbl 1486.14076 |
| [7] | Brubaker, M., Salzmann, M., Urtasun, R.: A family of MCMC methods on implicitly defined manifolds. In: Lawrence, N.D., Girolami, M., (eds.) Proceedings of the Fifteenth International Conference on Artificial Intelligence and Statistics, volume 22 of Proceedings of Machine Learning Research, La Palma, Canary Islands, pp. 161-172 (2012) |
| [8] | Cancès, E.; Legoll, F.; Stoltz, G., Theoretical and numerical comparison of some sampling methods for molecular dynamics, Math. Model. Numer. Anal., 41, 2, 351-389 (2007) · Zbl 1138.82341 · doi:10.1051/m2an:2007014 |
| [9] | Darve, E.; Chipot, C.; Pohorille, A., Thermodynamic integration using constrained and unconstrained dynamics, Free Energy Calculations, 119-170 (2007), Berlin: Springer, Berlin · doi:10.1007/978-3-540-38448-9_4 |
| [10] | Diaconis, P.; Holmes, S.; Shahshahani, M., Sampling from a manifold, Adv. Mod. Stat. Theory Appl., 10, 102-125 (2013) · Zbl 1356.62015 |
| [11] | Durmus, A., Moulines, E., Saksman, E.: On the convergence of Hamiltonian Monte Carlo. arXiv preprint arXiv:1705.00166 (2017) · Zbl 1460.65005 |
| [12] | Faou, E.; Lelièvre, T., Conservative stochastic differential equations: mathematical and numerical analysis, Math. Comput., 78, 2047-2074 (2009) · Zbl 1203.60069 · doi:10.1090/S0025-5718-09-02220-0 |
| [13] | Girolami, M.; Calderhead, B., Riemann manifold Langevin and Hamiltonian Monte Carlo methods, J. R. Stat. Soc. Ser. B (Methodol.), 73, 2, 1-37 (2011) · Zbl 1411.62071 |
| [14] | Hairer, E.; Lubich, C.; Wanner, G., Geometric Numerical Integration: Structure-Preserving Algorithms for Ordinary Differential Equations (2006), Berlin: Springer, Berlin · Zbl 1094.65125 |
| [15] | Hartmann, C., An ergodic sampling scheme for constrained Hamiltonian systems with applications to molecular dynamics, J. Stat. Phys., 130, 4, 687-711 (2008) · Zbl 1214.82063 · doi:10.1007/s10955-007-9470-2 |
| [16] | Hastings, Wk, Monte Carlo sampling methods using Markov chains and their applications, Biometrika, 57, 97-109 (1970) · Zbl 0219.65008 · doi:10.1093/biomet/57.1.97 |
| [17] | Horowitz, Am, A generalized guided Monte-Carlo algorithm, Phys. Lett. B, 268, 247-252 (1991) · doi:10.1016/0370-2693(91)90812-5 |
| [18] | Kaufman, Dm; Pai, Dk, Geometric numerical integration of inequality constrained, nonsmooth Hamiltonian systems, SIAM J. Sci. Comput., 34, 5, A2670-A2703 (2012) · Zbl 1259.65202 · doi:10.1137/100800105 |
| [19] | Leimkuhler, B.; Matthews, C., Efficient molecular dynamics using geodesic integration and solvent-solute splitting, Proc. R. Soc. A, 472, 20160138 (2016) · Zbl 1371.82078 · doi:10.1098/rspa.2016.0138 |
| [20] | Leimkuhler, B.; Reich, S., Symplectic integration of constrained Hamiltonian systems, Math. Comput., 63, 208, 589-605 (1994) · Zbl 0813.65103 · doi:10.1090/S0025-5718-1994-1250772-7 |
| [21] | Leimkuhler, B.; Reich, S., Simulating Hamiltonian Dynamics (2004), Cambridge: Cambridge University Press, Cambridge · Zbl 1069.65139 |
| [22] | Lelièvre, T.; Rousset, M.; Stoltz, G., Free-energy Computations: A Mathematical Perspective (2010), London: Imperial College Press, London · Zbl 1227.82002 · doi:10.1142/p579 |
| [23] | Lelièvre, T.; Rousset, M.; Stoltz, G., Langevin dynamics with constraints and computation of free energy differences, Math. Comput., 81, 2071-2125 (2012) · Zbl 1274.82034 · doi:10.1090/S0025-5718-2012-02594-4 |
| [24] | Livingstone, S., Betancourt, M., Byrne, S., Girolami, M.: On the geometric ergodicity of Hamiltonian Monte Carlo. arXiv preprint arXiv:1601.08057 (2016) · Zbl 1380.60070 |
| [25] | Marin, J-M; Pudlo, P.; Robert, C.; Ryder, R., Approximate Bayesian computational methods, Stat. Comput., 22, 1167-1180 (2012) · Zbl 1252.62022 · doi:10.1007/s11222-011-9288-2 |
| [26] | Mehlig, B.; Heermann, Dw; Forrest, Bm, Hybrid Monte Carlo method for condensed-matter systems, Phys. Rev. B, 45, 2, 679 (1992) · doi:10.1103/PhysRevB.45.679 |
| [27] | Metropolis, N.; Rosenbluth, Aw; Rosenbluth, Mn; Teller, Ah; Teller, E., Equations of state calculations by fast computing machines, J. Chem. Phys., 21, 6, 1087-1091 (1953) · Zbl 1431.65006 · doi:10.1063/1.1699114 |
| [28] | Rapaport, Dc, The Art of Molecular Dynamics Simulations (1995), Cambridge: Cambridge University Press, Cambridge |
| [29] | Roberts, Go; Rosenthal, Js, Optimal scaling of discrete approximations to Langevin diffusions, J. R. Stat. Soc. B, 60, 255-268 (1998) · Zbl 0913.60060 · doi:10.1111/1467-9868.00123 |
| [30] | Roberts, Go; Tweedie, Rl, Exponential convergence of Langevin distributions and their discrete approximations, Bernoulli, 2, 4, 341-363 (1996) · Zbl 0870.60027 · doi:10.2307/3318418 |
| [31] | Rossky, Pj; Doll, Jd; Friedman, Hl, Brownian dynamics as smart Monte Carlo simulation, J. Chem. Phys., 69, 10, 4628-4633 (1978) · doi:10.1063/1.436415 |
| [32] | Schütte, C.: Conformational dynamics: modelling, theory, algorithm and application to biomolecules. Habilitation dissertation, Free University Berlin (1998) |
| [33] | Schwartz, L.: Analyse I. Théorie des ensembles et topologie. Hermann, Paris (1991) · Zbl 0872.54001 |
| [34] | Stoltz, G.; Trstanova, Z., Stable and accurate schemes for Langevin dynamics with general kinetic energies, Multiscale Model. Simul., 16, 2, 777-806 (2018) · Zbl 1392.82004 · doi:10.1137/16M110575X |
| [35] | Tavaré, S.; Balding, D.; Griffith, R.; Donnelly, P., Inferring coalescence times from DNA sequence data, Genetics, 145, 505-518 (1997) |
| [36] | Zappa, E.; Holmes-Cerfon, M.; Goodman, J., Monte Carlo on manifolds: sampling densities and integrating functions, Commun. Pure Appl. Math., 71, 2609-2647 (2018) · Zbl 06993640 · doi:10.1002/cpa.21783 |
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.
