Abstract
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.
Similar content being viewed by others
Change history
22 November 2019
Shiva Darshan and Miranda Holmes–Cerfon (Courant Institute, NYU) pointed out a mistake in the projection functions to enforce the momentum constraint when rewriting the algorithm in Numerical Algorithm A of Section 3.1. Two different projection functions are actually needed, see indeed the formula for the Lagrange multiplier
22 November 2019
Shiva Darshan and Miranda Holmes���Cerfon (Courant Institute, NYU) pointed out a mistake in the projection functions to enforce the momentum constraint when rewriting the algorithm in Numerical Algorithm��A of Section��3.1. Two different projection functions are actually needed, see indeed the formula for the Lagrange multiplier
Notes
This scheme is now named “Hamiltonian Monte Carlo” in the computational statistics literature.
Notice that the composition of two Markov kernels which are \(\pi \)-reversible is not \(\pi \)-reversible in general; it however admits \(\pi \) as an invariant measure.
References
Abraham, R., Marsden, J.E.: Foundations of Mechanics. Advanced Book Program. Benjamin/Cummings Publishing Co. Inc., San Francisco (1978)
Afshar, H.M., Domke, J.: Reflection, refraction, and Hamiltonian Monte Carlo. In: Advances in Neural Information Processing Systems, pp. 3007–3015 (2015)
Andersen, H.C.: Rattle: a “velocity” version of the Shake algorithm for molecular dynamics calculations. J. Comput. Phys. 52, 24–34 (1983)
Arnol’d, V.I.: Mathematical Methods of Classical Mechanics. Graduate Texts in Mathematics, vol. 60. Springer, Berlin (1989)
Bou-Rabee, N., Vanden-Eijnden, E.: Pathwise accuracy and ergodicity of metropolized integrators for SDEs. Commun. Pure Appl. Math. 63(5), 655–696 (2009)
Breiding, P., Marigliano, O.: Sampling from the uniform distribution on an algebraic manifold. arXiv preprint arXiv:1810.06271 (2018)
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)
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)
Darve, E.: Thermodynamic integration using constrained and unconstrained dynamics. In: Chipot, C., Pohorille, A. (eds.) Free Energy Calculations, pp. 119–170. Springer, Berlin (2007)
Diaconis, P., Holmes, S., Shahshahani, M.: Sampling from a manifold. Adv. Mod. Stat. Theory Appl. 10, 102–125 (2013)
Durmus, A., Moulines, E., Saksman, E.: On the convergence of Hamiltonian Monte Carlo. arXiv preprint arXiv:1705.00166 (2017)
Faou, E., Lelièvre, T.: Conservative stochastic differential equations: mathematical and numerical analysis. Math. Comput. 78, 2047–2074 (2009)
Girolami, M., Calderhead, B.: Riemann manifold Langevin and Hamiltonian Monte Carlo methods. J. R. Stat. Soc. Ser. B (Methodol.) 73(2), 1–37 (2011)
Hairer, E., Lubich, C., Wanner, G.: Geometric Numerical Integration: Structure-Preserving Algorithms for Ordinary Differential Equations. Springer Series in Computational Mathematics, vol. 31. Springer, Berlin (2006)
Hartmann, C.: An ergodic sampling scheme for constrained Hamiltonian systems with applications to molecular dynamics. J. Stat. Phys. 130(4), 687–711 (2008)
Hastings, W.K.: Monte Carlo sampling methods using Markov chains and their applications. Biometrika 57, 97–109 (1970)
Horowitz, A.M.: A generalized guided Monte-Carlo algorithm. Phys. Lett. B 268, 247–252 (1991)
Kaufman, D.M., Pai, D.K.: Geometric numerical integration of inequality constrained, nonsmooth Hamiltonian systems. SIAM J. Sci. Comput. 34(5), A2670–A2703 (2012)
Leimkuhler, B., Matthews, C.: Efficient molecular dynamics using geodesic integration and solvent-solute splitting. Proc. R. Soc. A 472, 20160138 (2016)
Leimkuhler, B., Reich, S.: Symplectic integration of constrained Hamiltonian systems. Math. Comput. 63(208), 589–605 (1994)
Leimkuhler, B., Reich, S.: Simulating Hamiltonian Dynamics. Cambridge University Press, Cambridge (2004)
Lelièvre, T., Rousset, M., Stoltz, G.: Free-energy Computations: A Mathematical Perspective. Imperial College Press, London (2010)
Lelièvre, T., Rousset, M., Stoltz, G.: Langevin dynamics with constraints and computation of free energy differences. Math. Comput. 81, 2071–2125 (2012)
Livingstone, S., Betancourt, M., Byrne, S., Girolami, M.: On the geometric ergodicity of Hamiltonian Monte Carlo. arXiv preprint arXiv:1601.08057 (2016)
Marin, J.-M., Pudlo, P., Robert, C., Ryder, R.: Approximate Bayesian computational methods. Stat. Comput. 22, 1167–1180 (2012)
Mehlig, B., Heermann, D.W., Forrest, B.M.: Hybrid Monte Carlo method for condensed-matter systems. Phys. Rev. B 45(2), 679 (1992)
Metropolis, N., Rosenbluth, A.W., Rosenbluth, M.N., Teller, A.H., Teller, E.: Equations of state calculations by fast computing machines. J. Chem. Phys. 21(6), 1087–1091 (1953)
Rapaport, D.C.: The Art of Molecular Dynamics Simulations. Cambridge University Press, Cambridge (1995)
Roberts, G.O., Rosenthal, J.S.: Optimal scaling of discrete approximations to Langevin diffusions. J. R. Stat. Soc. B 60, 255–268 (1998)
Roberts, G.O., Tweedie, R.L.: Exponential convergence of Langevin distributions and their discrete approximations. Bernoulli 2(4), 341–363 (1996)
Rossky, P.J., Doll, J.D., Friedman, H.L.: Brownian dynamics as smart Monte Carlo simulation. J. Chem. Phys. 69(10), 4628–4633 (1978)
Schütte, C.: Conformational dynamics: modelling, theory, algorithm and application to biomolecules. Habilitation dissertation, Free University Berlin (1998)
Schwartz, L.: Analyse I. Théorie des ensembles et topologie. Hermann, Paris (1991)
Stoltz, G., Trstanova, Z.: Stable and accurate schemes for Langevin dynamics with general kinetic energies. Multiscale Model. Simul. 16(2), 777–806 (2018)
Tavaré, S., Balding, D., Griffith, R., Donnelly, P.: Inferring coalescence times from DNA sequence data. Genetics 145, 505–518 (1997)
Zappa, E., Holmes-Cerfon, M., Goodman, J.: Monte Carlo on manifolds: sampling densities and integrating functions. Commun. Pure Appl. Math. 71, 2609–2647 (2018)
Acknowledgements
We thank Christian Robert for pointing out the reference [36] as soon as it was preprinted on arXiv, as well as Jonathan Goodman and Miranda Cerfon–Holmes for very useful discussions. This work was funded in part by the European Research Council under the European Union’s Seventh Framework Programme (FP/2007-2013) / ERC Grant Agreement number 614492, as well as the Agence Nationale de la Recherche, under grant ANR-14-CE23-0012 (COSMOS). We also benefited from the scientific environment of the Laboratoire International Associé between the Centre National de la Recherche Scientifique and the University of Illinois at Urbana-Champaign.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Lelièvre, T., Rousset, M. & Stoltz, G. Hybrid Monte Carlo methods for sampling probability measures on submanifolds. Numer. Math. 143, 379–421 (2019). https://doi.org/10.1007/s00211-019-01056-4
Received:
Revised:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00211-019-01056-4