Langevin diffusion for population based sampling with an application in Bayesian inference for pharmacodynamics. (English) Zbl 1390.62031
Summary: We propose an algorithm for the efficient and robust sampling of the posterior probability distribution in Bayesian inference problems. The algorithm combines the local search capabilities of the manifold Metropolis adjusted Langevin transition kernels with the advantages of global exploration by a population based sampling algorithm, the transitional Markov chain Monte Carlo (TMCMC). The Langevin diffusion process is determined by either the Hessian or the Fisher information of the target distribution with appropriate modifications for non-positive definiteness. The present method is shown to be superior to other population based algorithms, in sampling probability distributions for which gradients are available, and is shown to handle otherwise unidentifiable models. We demonstrate the capabilities and advantages of the method in computing the posterior distribution of the parameters in a Pharmacodynamics model, for glioma growth and its drug induced inhibition, using clinical data.
MSC:
| 62F15 | Bayesian inference |
| 62D05 | Sampling theory, sample surveys |
| 62P10 | Applications of statistics to biology and medical sciences; meta analysis |
References:
| [1] | K. Aizu, {\it Parameter differentiation of quantum-mechanical linear operators}, J. Math. Phys., 4 (1963), pp. 762-775. · Zbl 0122.22301 |
| [2] | J. L. Beck and K. V. Yuen, {\it Model selection using response measurements: Bayesian probabilistic approach}, J. Engrg. Mech., 130 (2004), pp. 192-203. |
| [3] | A. Beskos, A. Jasra, K. Law, Y. Marzouk, and Y. Zhou, {\it Multilevel Sequential Monte Carlo with Dimension-Independent Likelihood-Informed Proposals}, preprint, , 2017. · Zbl 1395.82229 |
| [4] | M. Betancourt, {\it A general metric for Riemannian manifold Hamiltonian Monte Carlo}, in Proceedings of the 1st International Conference on Geometric Science of Information, GSI 2013 (Paris, France), Lecture Notes in Comput. Sci. 8085, Springer, Berlin, Heidelberg, 2013, pp. 327-334. · Zbl 1406.65004 |
| [5] | W. Betz, I. Papaioannou, and D. Straub, {\it Transitional Markov chain Monte Carlo: Observations and improvements}, J. Engrg. Mech., 142 (2016), 04016016. |
| [6] | M. A. Brubaker, M. Salzmann, and R. Urtasun, {\it A family of MCMC methods on implicitly defined manifolds}, in Proceedings of the 15th International Conference on Artificial Intelligence and Statistics (AISTATS-12), Proc. Machine Learn. Res. 22, PMLR, Cambridge, MA, 2012, pp. 161-172. |
| [7] | T. Bui-Thanh and O. Ghattas, {\it A Scaled Stochastic Newton Algorithm for Markov Chain Monte Carlo Simulations}, manuscript, 2012. |
| [8] | B. Calderhead and M. Girolami, {\it Estimating Bayes factors via thermodynamic integration and population MCMC}, Comput. Statist. Data Anal., 53 (2009), pp. 4028-4045. · Zbl 1453.62055 |
| [9] | M. Chiachio, J. L. Beck, J. Chiachio, and G. Rus, {\it Approximate Bayesian computation by subset simulation}, SIAM J. Sci. Comput., 36 (2014), pp. A1339-A1358, . · Zbl 1297.65013 |
| [10] | J. Ching and Y.-C. Chen, {\it Transitional Markov chain Monte Carlo method for Bayesian model updating, model class selection, and model averaging}, J. Engrg. Mech., 133 (2007), pp. 816-832. |
| [11] | T. Cui, K. J. H. Law, and Y. M. Marzouk, {\it Dimension-independent likelihood-informed MCMC}, J. Comput. Phys., 304 (2016), pp. 109-137. · Zbl 1349.65009 |
| [12] | T. Cui, J. Martin, Y. M. Marzouk, A. Solonen, and A. Spantini, {\it Likelihood-informed dimension reduction for nonlinear inverse problems}, Inverse Problems, 30 (2014), 114015. · Zbl 1310.62030 |
| [13] | P. Del Moral, A. Doucet, and A. Jasra, {\it Sequential Monte Carlo samplers}, J. R. Stat. Soc. Ser. B Stat. Methodol., 68 (2006), pp. 411-436. · Zbl 1105.62034 |
| [14] | S. Duane, A. Kennedy, B. J. Pendleton, and D. Roweth, {\it Hybrid Monte Carlo}, Phys. Lett. B, 195 (1987), pp. 216-222. |
| [15] | N. Friel and A. N. Pettitt, {\it Marginal likelihood estimation via power posteriors}, J. R. Stat. Soc. Ser. B Stat. Methodol., 70 (2008), pp. 589-607. · Zbl 05563360 |
| [16] | M. Girolami and B. Calderhead, {\it Riemann manifold Langevin and Hamiltonian Monte Carlo methods}, J. R. Stat. Soc. Ser. B Stat. Methodol., 73 (2011), pp. 123-214. · Zbl 1411.62071 |
| [17] | P. J. Green, K. Łatuszyński, M. Pereyra, and C. P. Robert, {\it Bayesian computation: A summary of the current state, and samples backwards and forwards}, Stat. Comput., 25 (2015), pp. 835-862. · Zbl 1331.62017 |
| [18] | H. Haario, M. Laine, A. Mira, and E. Saksman, {\it DRAM: Efficient adaptive MCMC}, Stat. Comput., 16 (2006), pp. 339-354. |
| [19] | P. Hadjidoukas, P. Angelikopoulos, C. Papadimitriou, and P. Koumoutsakos, {\it \(Π 4\) U: A high performance computing framework for Bayesian uncertainty quantification of complex models}, J. Comput. Phys., 284 (2015), pp. 1-21. · Zbl 1352.65009 |
| [20] | N. Hansen, S. D. Müller, and P. Koumoutsakos, {\it Reducing the time complexity of the derandomized evolution strategy with covariance matrix adaptation (CMA-ES)}, Evolutionary Comput., 11 (2003), pp. 1-18. |
| [21] | K. Hukushima and K. Nemoto, {\it Exchange Monte Carlo method and application to spin glass simulations}, J. Phys. Soc. Japan, 65 (1996), pp. 1604-1608. |
| [22] | J. K. Ghosh, M. Delampady, and T. Samanta, {\it An Introduction to Bayesian Analysis}, Springer, New York, 2006. · Zbl 1135.62002 |
| [23] | E. T. Jaynes, {\it Probability Theory: The Logic of Science}, Cambridge University Press, Cambridge, UK, 2003. · Zbl 1045.62001 |
| [24] | M. C. Kennedy and A. O’Hagan, {\it Bayesian calibration of computer models}, J. R. Stat. Soc. Ser. B Stat. Methodol., 63 (2001), pp. 425-464. · Zbl 1007.62021 |
| [25] | T. S. Kleppe, {\it Adaptive step size selection for Hessian-based manifold Langevin samplers}, Scand. J. Stat., 43 (2016), pp. 788-805. · Zbl 1468.62435 |
| [26] | S. C. Kou, Q. Zhou, and W. H. Wong, {\it Equi-energy sampler with applications in statistical inference and statistical mechanics}, Ann. Statist., 34 (2006), pp. 1581-1652. · Zbl 1246.82054 |
| [27] | N. Lartillot and H. Philippe, {\it Computing Bayes factors using thermodynamic integration}, System. Biol., 55 (2006), pp. 195-207. |
| [28] | P. Marjoram, J. Molitor, V. Plagnol, and S. Tavaré, {\it Markov chain Monte Carlo without likelihoods}, Proc. Natl. Acad. Sci. USA, 100 (2003), pp. 15324-15328. |
| [29] | J. Martin, L. C. Wilcox, C. Burstedde, and O. Ghattas, {\it A stochastic Newton MCMC method for large-scale statistical inverse problems with application to seismic inversion}, SIAM J. Sci. Comput., 34 (2012), pp. A1460-A1487, . · Zbl 1250.65011 |
| [30] | R. M. Neal, {\it Annealed Importance Sampling}, preprint, , 1998. |
| [31] | A. Raue, C. Kreutz, F. J. Theis, and J. Timmer, {\it Joining forces of Bayesian and frequentist methodology: A study for inference in the presence of non-identifiability}, Philos. Trans. R. Soc. Lond. Ser. A Math. Phys. Eng. Sci., 371 (2012), 20110544. · Zbl 1353.62013 |
| [32] | B. Ribba, G. Kaloshi, M. Peyre, D. Ricard, V. Calvez, M. Tod, B. Čajavec-Bernard, A. Idbaih, D. Psimaras, L. Dainese, J. Pallud, S. Cartalat-Carel, J. Y. Delattre, J. Honnorat, E. Grenier, and F. Ducray, {\it A tumor growth inhibition model for low-grade glioma treated with chemotherapy or radiotherapy}, Clinical Cancer Res., 18 (2012), pp. 5071-5080. |
| [33] | C. P. Robert and G. Casella, {\it Monte Carlo Statistical Methods}, 2nd ed., Springer Texts Statist., Springer-Verlag New York, 2004. · Zbl 1096.62003 |
| [34] | G. Roberts and O. Stramer, {\it Langevin diffusions and Metropolis-Hastings algorithms}, Methodol. Comput. Appl. Probab., 4 (2002), pp. 337-357. · Zbl 1033.65003 |
| [35] | G. O. Roberts and J. S. Rosenthal, {\it Optimal scaling of discrete approximations to Langevin diffusions}, J. R. Stat. Soc. Ser. B Stat. Methodol., 60 (1998), pp. 255-268. · Zbl 0913.60060 |
| [36] | G. O. Roberts and J. S. Rosenthal, {\it Optimal scaling for various Metropolis-Hastings algorithms}, Stat. Sci., 16 (2001), pp. 351-367. · Zbl 1127.65305 |
| [37] | R. Sherman, K. Davies, M. Robb, N. Hunter, and R. Califf, {\it Accelerating development of scientific evidence for medical products within the existing US regulatory framework}, Nature Rev. Drug Discov., 16 (2017), pp. 297-298. |
| [38] | M. Slotani, {\it Tolerance regions for a multivariate normal population}, Ann. Inst. Statist. Math., 16 (1964), pp. 135-153. · Zbl 0124.09605 |
| [39] | A. Solonen, P. Ollinaho, M. Laine, H. Haario, J. Tamminen, and H. Järvinen, {\it Efficient MCMC for climate model parameter estimation: Parallel adaptive chains and early rejection}, Bayesian Anal., 7 (2012), pp. 715-736. · Zbl 1330.60091 |
| [40] | S. Wu, P. Angelikopoulos, C. Papadimitriou, and P. Koumoutsakos, {\it Bayesian Annealed Sequential Importance Sampling (BASIS): An unbiased version of Transitional Markov Chain Monte Carlo}, ASCE-ASME J. Risk Uncertainty Engrg. Syst. B Mech. Engrg., 4 (2017), RISK-16-1101. |
| [41] | T. Xifara, C. Sherlock, S. Livingstone, S. Byrne, and M. Girolami, {\it Langevin diffusions and the Metropolis-adjusted Langevin algorithm}, Statist. Probab. Lett., 91 (2014), pp. 14-19. · Zbl 1298.60081 |
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.
