A Hamiltonian Monte Carlo Method for Non-Smooth Energy Sampling

Efficient sampling from high-dimensional distributions is a challenging issue that is encountered in many large data recovery problems. In this context, sampling using Hamiltonian dynamics is one of the recent techniques that have been proposed to exploit the target distribution geometry. Such schemes have clearly been shown to be efficient for multidimensional sampling but, rather, are adapted to distributions from the exponential family with smooth energy functions. In this paper, we address the problem of using Hamiltonian dynamics to sample from probability distributions having non-differentiable energy functions such as those based on the l1 norm. Such distributions are being used intensively in sparse signal and image recovery applications. The technique studied in this paper uses a modified leapfrog transform involving a proximal step. The resulting nonsmooth Hamiltonian Monte Carlo method is tested and validated on a number of experiments. Results show its ability to accurately sample according to various multivariate target distributions. The proposed technique is illustrated on synthetic examples and is applied to an image denoising problem.

[1]  Edward H. Adelson,et al.  Noise removal via Bayesian wavelet coring , 1996, Proceedings of 3rd IEEE International Conference on Image Processing.

[2]  Valérie R. Wajs,et al.  A variational formulation for frame-based inverse problems , 2007 .

[3]  Jérôme Idier,et al.  2 S ep 2 01 4 Efficient Gaussian Sampling for Solving Large-Scale Inverse Problems using MCMC Methods ∗ , .

[4]  Patrick L. Combettes,et al.  Iterative image deconvolution using overcomplete representations , 2006, 2006 14th European Signal Processing Conference.

[5]  Radford M. Neal.  MCMC Using Hamiltonian Dynamics , 2011, 1206.1901.

[6]  Patrick L. Combettes,et al.  Signal Recovery by Proximal Forward-Backward Splitting , 2005, Multiscale Model. Simul..

[7]  Alfred O. Hero,et al.  Hierarchical Bayesian Sparse Image Reconstruction With Application to MRFM , 2008, IEEE Transactions on Image Processing.

[8]  J. Richard,et al.  Efficient high-dimensional importance sampling , 2007 .

[9]  Boualem Boashash,et al.  Wavelet Denoising Based on the MAP Estimation Using the BKF Prior With Application to Images and EEG Signals , 2013, IEEE Transactions on Signal Processing.

[10]  Kenneth M. Hanson,et al.  Markov chain Monte Carlo posterior sampling with the Hamiltonian method , 2001, SPIE Medical Imaging.

[11]  Ryan P. Adams,et al.  Elliptical slice sampling , 2009, AISTATS.

[12]  Patrick L. Combettes,et al.  Proximal Splitting Methods in Signal Processing , 2009, Fixed-Point Algorithms for Inverse Problems in Science and Engineering.

[13]  H. H. Rachford,et al.  On the numerical solution of heat conduction problems in two and three space variables , 1956 .

[14]  Amel Benazza-Benyahia,et al.  A hierarchical Bayesian model for frame representation , 2010, 2010 IEEE International Conference on Acoustics, Speech and Signal Processing.

[15]  J. Moreau.  Proximité et dualité dans un espace hilbertien , 1965 .

[16]  Saïd Moussaoui,et al.  Efficient Gaussian Sampling for Solving Large-Scale Inverse Problems Using MCMC , 2014, IEEE Transactions on Signal Processing.

[17]  Christian P. Robert,et al.  Monte Carlo Statistical Methods , 2005, Springer Texts in Statistics.

[18]  R. Tweedie,et al.  Exponential convergence of Langevin distributions and their discrete approximations , 1996 .

[19]  É. Moulines,et al.  On stochastic proximal gradient algorithms , 2014 .

[20]  F. Clarke.  Optimization And Nonsmooth Analysis , 1983 .

[21]  Jean-Yves Tourneret,et al.  Sparse Bayesian regularization using Bernoulli-Laplacian priors , 2013, 21st European Signal Processing Conference (EUSIPCO 2013).

[22]  Jean-Yves Tourneret,et al.  Unsupervised Post-Nonlinear Unmixing of Hyperspectral Images Using a Hamiltonian Monte Carlo Algorithm , 2014, IEEE Transactions on Image Processing.

[23]  Jean-Yves Tourneret,et al.  Sparse signal recovery using a Bernoulli generalized Gaussian prior , 2015, 2015 23rd European Signal Processing Conference (EUSIPCO).

[24]  Eero P. Simoncelli,et al.  Image quality assessment: from error visibility to structural similarity , 2004, IEEE Transactions on Image Processing.

[25]  Jean-Yves Tourneret,et al.  Sparse EEG Source Localization Using Bernoulli Laplacian Priors , 2015, IEEE Transactions on Biomedical Engineering.

[26]  Brian D. Jeffs,et al.  Adaptive image restoration using a generalized Gaussian model for unknown noise , 1995, IEEE Trans. Image Process..

[27]  Florian Steinke,et al.  Bayesian Inference and Optimal Design in the Sparse Linear Model , 2007, AISTATS.

[28]  P. Moulin,et al.  Analysis of multiresolution image denoising schemes using generalized-Gaussian priors , 1998, Proceedings of the IEEE-SP International Symposium on Time-Frequency and Time-Scale Analysis (Cat. No.98TH8380).

[29]  Christophe Croux,et al.  An Information Criterion for Variable Selection in Support Vector Machines , 2007 .

[30]  Ari Pakman,et al.  Exact Hamiltonian Monte Carlo for Truncated Multivariate Gaussians , 2012, 1208.4118.

[31]  George Eastman House,et al.  Sparse Bayesian Learning and the Relevan e Ve tor Ma hine , 2001 .

[32]  S. Frick,et al.  Compressed Sensing , 2014, Computer Vision, A Reference Guide.

[33]  Nando de Freitas,et al.  Adaptive Hamiltonian and Riemann Manifold Monte Carlo , 2013, ICML.

[34]  Alfred O. Hero,et al.  A Survey of Stochastic Simulation and Optimization Methods in Signal Processing , 2015, IEEE Journal of Selected Topics in Signal Processing.

[35]  V. Jeyakumar,et al.  Approximate Jacobian Matrices for Nonsmooth Continuous Maps and C 1 -Optimization , 1998 .

[36]  Jean-François Giovannelli,et al.  Sampling High-Dimensional Gaussian Distributions for General Linear Inverse Problems , 2012, IEEE Signal Processing Letters.

[37]  Andrew Gelman,et al.  Handbook of Markov Chain Monte Carlo , 2011 .

[38]  O HeroAlfred,et al.  Hierarchical Bayesian sparse image reconstruction with application to MRFM , 2009 .

[39]  M. Girolami,et al.  Riemann manifold Langevin and Hamiltonian Monte Carlo methods , 2011, Journal of the Royal Statistical Society: Series B (Statistical Methodology).

[40]  B. Alder,et al.  Studies in Molecular Dynamics. I. General Method , 1959 .

[41]  P. L. Combettes,et al.  A proximal decomposition method for solving convex variational inverse problems , 2008, 0807.2617.

[42]  Marcelo Pereyra,et al.  Proximal Markov chain Monte Carlo algorithms , 2013, Statistics and Computing.

[43]  Amel Benazza-Benyahia,et al.  A wavelet-based regularized reconstruction algorithm for SENSE parallel MRI with applications to neuroimaging , 2011, Medical Image Anal..

[44]  Jean-Christophe Pesquet,et al.  Spatio-temporal wavelet regularization for parallel MRI reconstruction: application to functional MRI , 2011, Magnetic Resonance Materials in Physics, Biology and Medicine.

[45]  Heinz H. Bauschke,et al.  Convex Analysis and Monotone Operator Theory in Hilbert Spaces , 2011, CMS Books in Mathematics.