Transport maps and dimension reduction for Bayesian computation Youssef Marzouk

Size: px
Start display at page:

Download "Transport maps and dimension reduction for Bayesian computation Youssef Marzouk"

Transcription

1 Transport maps and dimension reduction for Bayesian computation Youssef Marzouk Massachusetts Institute of Technology Department of Aeronautics & Astronautics Center for Computational Engineering

2 Acknowledgments Joint work with: Matt Parno, Tarek Moselhy (MIT) Alessio Spantini, Tiangang Cui, Antti Solonen (MIT); Kody Law (KAUST); James Martin (UT-Austin); Luis Tenorio (Mines) Support from DOE Applied Math program (ASCR)

3 Inverse problems Example: subsurface flow and transport forward model (PDE) d =G(θ)+ observation/model errors Data d are typically limited, noisy, and indirect Forward model G may be computationally intensive Parameter θ is often high- or infinite-dimensional

4 Inverse problems Example: ice-ocean interactions o S o W o W o W 30 Pine Island Glacier (left) modeled with the MIT General Circulation Model (right) [P. Conrad, P. Heimbach, YM]

5 UQ for inverse problems z } { z} { We adopt a Bayesian approach: p(θ d) p(d θ)p(θ) {z } {z } Key idea: model parameters θ are treated as random variables

6 Computational challenges Extract information from the posterior by evaluating posterior expectations: E π [f ] = f (θ)π(θ)dθ Key strategies for making this computationally tractable: Approximations of the forward model: spectral expansions, local interpolants, reduced order models, multi-fidelity approaches Efficient and structure-exploiting sampling schemes to explore the posterior distribution

7 Computational challenges Extract information from the posterior by evaluating posterior expectations: E π [f ] = f (θ)π(θ)dθ Key strategies for making this computationally tractable: Approximations of the forward model: spectral expansions, local interpolants, reduced order models, multi-fidelity approaches Efficient and structure-exploiting sampling schemes to explore the posterior distribution

8 Outline 1. Transport maps for Bayesian computation 2. Dimension reduction and dimension-independent posterior sampling

9 Posterior sampling Markov chain Monte Carlo (MCMC) algorithms are the workhorse of Bayesian computation Metropolis-Hastings kernel K(θ t, ) 1. Propose a sample θ ~ q(θ θ t ) 2. Compute acceptance probability: α( θ,θ t ) = min 1, π( θ )q(θ t θ ) π(θ t )q( θ θ t ) 3. Update the chain state: θ t+1 = θ with probability α θ t with probability 1 α Good mixing requires design of effective proposal distributions q( θ θ)

10 Posterior sampling Markov chain Monte Carlo (MCMC) algorithms are the workhorse of Bayesian computation Effective = adapted to the target Can we transform proposals or targets for better sampling?

11 Optimal transport Deterministic coupling of two random variables, r ~ µ, Z ~ ν Z T T(r) r r Monge problem: min T c ( r,t(r) ) µ(dr), where T µ = ν A unique and monotone solution exists for quadratic (and other) transport costs c(x,y) [Brenier 1991, McCann 1995]

12 Transport maps: computation Alternatives to Monge problem: seek triangular (Knothe- Rosenblatt) maps Previous work: directly finding a map from prior to posterior [Moselhy & YM, JCP 2012] Reference = prior or an uncorrelated multivariate normal Target = posterior max E T T p log( π(t(r) )+ log det T r Map has exploitable structure in high-dimensional inverse problems: In an appropriate basis, map can be expressed as a low-rank perturbation of the identity output index input index

13 Combining transport maps with MCMC Previous optimization problem can be costly Map must be represented in a finite basis (e.g., polynomials) and thus may be approximate. Can we still achieve exact posterior sampling? Key idea: combining map construction with MCMC Posterior sampling + convex optimization Transport map preconditions MCMC sampling; posterior samples enable simpler map construction Can also be understood in the framework of adaptive MCMC

14 Constructing a map from samples Candidate map T yields approximation π of target distribution Optimization objective: min T max T D ( KL π(θ) π(θ; T) ) E π log p( T(θ))+ log D θt(θ) ( ) T ( ) Now π is the distribution on the map inputs and p is the distribution on the outputs Samples from π approximate the expectation; p has useful structure ( ) r p(r) = N(0,I)

15 Constructing a map from samples Additional structure: Seek a monotone triangular map (converges to Knothe-Rosenblatt rearrangement) T(θ) = T 1 (θ 1 ) T 2 (θ 1,θ 2 ) T D (θ 1,θ 2,,θ D ) Let target p(r) be multivariate normal with identity covariance Yields a convex and separable optimization problem: N 1 T j i=1 2 min T j 2 (θ (i) ) log T j θ j Sample-average approximation (SAA) with N samples from π Linear representation of map T (e.g., polynomial or RBF basis) θ (i ), s.t. T j θ j θ (i ) d min > 0, i {1,,N }

16 Map-accelerated MCMC Ingredient #1: static map Idea: perform MCMC in the reference space, on a preconditioned density Simple proposal in reference space (e.g., random walk) corresponds to a more complex/tailored proposal on target ( ) T ( ) p(r) r

17 Map-accelerated MCMC Ingredient #1: static map Idea: perform MCMC in the reference space, on a preconditioned density Simple proposal in reference space (e.g., random walk) corresponds to a more complex/tailored proposal on target ( ) T ( ) 0 r 0 r q r (r 0 r) r α = π( T 1 ( r )) D rt 1 q (r r r ) r π( T 1 (r)) D rt 1 r q r ( r r) simple proposal q r on pushforward of target through map

18 Map-accelerated MCMC Ingredient #1: static map Idea: perform MCMC in the reference space, on a preconditioned density Simple proposal in reference space (e.g., random walk) corresponds to a more complex/tailored proposal on target ( ) T ( ) 0 r 0 r q r (r 0 r) r α = π( T 1 ( r )) D rt 1 q (r r r ) r π( T 1 (r)) D rt 1 r q r ( r r) more complex proposal, directly on target distribution

19 Map-accelerated MCMC Ingredient #2: adaptive map Update the map with each MCMC iteration: more samples from π, more accurate E π, better T Analogous to adaptive MCMC [Haario 2001, Andrieu 2006] but with nonlinear transformation to capture non-gaussian structure ( ) T k ( ) p k (r) r

20 Map-accelerated MCMC Ingredient #2: adaptive map Update the map with each MCMC iteration: more samples from π, more accurate E π, better T Analogous to adaptive MCMC [Haario 2001, Andrieu 2006] but with nonlinear transformation to capture non-gaussian structure T k+1 ( ) ( ) p k+1 (r) r

21 Map-accelerated MCMC Inefficient; the changing correlation Ingredient #3: global proposals structure of limits Once the map becomes sufficiently accurate,mixing. would like to avoid random-walk behavior mapped random walk proposal reference random walk proposal qr (r 0 r ) = N(r, 2 I) q ( 0 ) = qr T ( 0) T ( ) detd T ( 0)

22 Map-accelerated MCMC Ingredient #3: global proposals Once the map becomes sufficiently accurate, would like to avoid random-walk behavior mapped independence proposal reference independence proposal 0 qr (r r ) = N(0, I ) q ( 0 ) = qr T ( 0) T ( ) detd T ( 0)

23 Map-accelerated MCMC Ingredient #3: global proposals Once the map becomes sufficiently accurate, would like to avoid random-walk behavior Solution: delayed rejection MCMC [Mira 2001] First proposal = independent sample from p (global, more efficient); second proposal = random walk (local, more robust) Entire scheme is provably ergodic with respect to the exact posterior measure [Parno, YM; arxiv: ] Requires enforcing a bi-lipschitz condition on maps, to preserve reasonable tail behavior of target

24 Lines in the reference space become geodesics on this manifold Example: BOD model Application 1: BOD inference problem Biochemical oxygen demand (BOD) model Goal: Infer parameters of a biochemical oxygen demand (BOD) model, given many early observations. Simple two-parameter inference problem: A simple BOD model, often used in water quality monitoring, is: B(t) =t )0) (1+ exp( B(ti ) = θ0 (1 exp( θ i 1 i 1t)). We assume 20 observations of B(t) are available in t 2of [1,B(t) 5]. Flatthatpriors; 20 noisy observations Final reference density p (r ) r2 2 Posterior distribution available for t [1,5] 1 r1

25 Example: BOD model r1 Comparison 1of sampling efficiency, for multiple MCMC schemes: pplication 1: BOD inference results minimum ESS/sec minimum ESS/evaluation TM+MIX 1.8e-1 TM+DRL 1.6e-1 NUTS 6.6e-3 AMALA DRAM 5.8e smmala 1.4 smmala 2.8e-3 DRAM TM+DRG NUTS 1.4e-3 AMALA TM+DRL 5.7e-2 TM+DRG TM+MIX ,000 2,000

26 NUTS AMALA Example: predator-prey model smmala a few initial observations DRAM given Infer ODE parameters 0 dp P 1,000PQ = rp 1 s dt K a +P 0.2 model 2,000 Population dq PQ =u vq a +P n a few initial population dt observations. Posterior predictive: Prey Population P(t) Predator Population Q(t) Time

27 Application 2: Predator-prey model Example: predator-prey model Population Goal: Infer ODE parameters given a few initial population observations. Posterior distribution: Posterio P0 Posterior distribution has complex structure Q0 r K s a u v dp = rp 1 dt dq PQ = u dt a+

28 Example: predator-prey model Table 5: Performance of MCMC samplers on predator-prey parameter inference problem. For this 8 dimensional Sampling problem, efficiency the maximum comparisons: correlation time and corresponding minimum e ective sample size is displayed provided. The relative performance compared to DRAM is also reported in the Rel. ESS/sec and Rel. ESS/eval columns. Method max ESS ESS/sec ESS/eval Rel. ESS/sec Rel. ESS/eval DRAM e e smmala e e AMALA e e TM+DRG e e TM+DRL e e TM+RWM e e TM+MIX e e TM+LA e e

29 Map-accelerated MCMC Important characteristics of the approach: Map construction is easily parallelizable (Currently) requires no gradients from posterior density Generalizes many current MCMC techniques: Adaptive Metropolis: map enables non-gaussian proposals; mixing local and global moves Manifold MCMC [Girolami & Calderhead 2011]: map defines a Riemannian metric; paths in reference space are geodesics on target What about high dimensions? How to choose or enrich the space of triangular maps T T?

30 Outline 1. Transport maps for Bayesian computation 2. Dimension reduction and dimension-independent posterior sampling

31 Dimension reduction Bayesian inverse problems are in principle infinitedimensional (practically high-dimensional) But what controls the intrinsic dimensionality? 1. Smoothing from the prior distribution (e.g., correlation) 2. How many observations are used 3. How much the forward model filters the parameters (ill-posedness of the classical/deterministic problem) More formally, #2 and #3 contribute to low rank of the data misfit Hessian H(x) = 2 x log p(d x) Key idea: low dimensionality lies in the change from prior to posterior

32 Low-dimensional structure Algorithm design: identify/use directions along which posterior is most changed relative to the prior Developments along this theme: Linear Bayesian inverse problem: optimality results, fast posterior mean estimators Nonlinear Bayesian inverse problem: approximate posterior, explicit dimension reduction Infinite-dimensional setting: dimension-independent + likelihood-informed (DILI) samplers

33 Linear Bayesian IP Simple linear-gaussian model: d =Gx +, x ~ N(0,Γ pr ), ~ N(0,Γ obs ); H :=G T Γ 1 obs G Exact posterior covariance is: Γ pos = Γ pr Γ pr G T Γ y 1 GΓ pr, with Γ d := Γ obs +GΓ pr G T Posterior mean is: µ pos = Γ pos G T Γ 1 obs d

34 Linear Bayesian IP Consider approximations of the form M r := ˆΓ pos = Γ pr KK T 0, rank(k) r { } = Γ pos Γ pr KK Theorem: optimal approximations of the posterior covariance [Spantini et al., arxiv: ] Optimality in Förstner-Moonen metric, rather than Frobenius Reinforces importance of prior-preconditioned Hessian [Flath, Ghattas, et al. 2011]

35 Linear Bayesian IP (Optimal covariance approximation) Let (δ 2 i,z i ) be the eigenvalue-eigenvector pairs of the priorpreconditioned Hessian Ĥ = Γ 1/2 1/2 pr H Γ pr. Then a ˆΓ that minimizes d ( F Γ pos, ˆΓ ) pos M r pos is r 2 Γ pos = Γ pr δ i (1 + δ 2 i ) 1 w i w T i, where w i = Γ pr i=1 1/2 z i d F is a Riemannian metric on the space of real symmetric positive definite matrices [Förstner & Moonen 2003] d 2 F (A,B) = tr ln 2 (A 1/2 BA 1/2 ) = Also yields optimal approximation of Gaussian posterior in Kullback- Leibler divergence and Hellinger metric i ln2 ( λ i (A,B))

36 Linear Bayesian IP Theorem: optimal approximations of the posterior mean [Spantini et al., arxiv: ] Consider approximations in the form of a low-rank matrix applied to the observations: { } ˆµ pos = Ad, A A r := A : rank(a r ) r Find A A r minimizing Bayes risk under squared-error loss: A 2 = arg min E A A x,d Ad x 1 Γpos r A Result: easily computed from low-rank approximation of Γ 1/2 1/2 obs G Γ pr

37 Linear Bayesian IP µ pos normalized error order of approximating class

38 Nonlinear Bayesian IP Nonlinear forward model: Hessian depends on parameter Find a linear subspace that captures likelihood-informed directions; use posterior expectation of pp-hessian Enables: π(x d) π(x d) = p(y Γ 1/2 pr U r x r )p(x r ) p(x ) = π r (x r d) π (x ) Accelerated MCMC in lower dimensions Variance reduction via Rao-Blackwellization x = Γ 1/2 pr U r x r + Γ 1/2 pr (I U r U T r )x ; x ~ N(0,I ) from MCMC samples [Cui et al., Inverse Problems 2014] everything known analytically!

39 Nonlinear Bayesian IP Posterior variance: elliptic inverse problem (4800 dim) (a) LIS (b) CS s s s 1 (c) LIS + CS s 1 (d) full-space MCMC s s s 1 s 1

40 DILI samplers What about exact sampling in the function space setting? MH kernel: ν(du,d u ) = q(u,d u )µ(du) ν (du,d u ) = q( u,du)µ(d u ) α(u k, u ) = min 1, dν dν (u, u ) k Need ν ν to define a valid acceptance probability How? Use discretizations of Langevin SDE that exactly preserve a Gaussian measure dominating the posterior Dimension-independent + likelihood-informed (DILI): A general class of operator-weighted MCMC proposals on function space Operators encode LIS structure; finite + infinite dimensions [Cui, Law, YM; arxiv: ]

41 DILI samplers Sampling efficiency, elliptic PDE inverse problem (n=1600) LI improves mixing significantly over pcn DI and/or particular choice of LIS improve it further 1 (c) autocorrelations 1 (c) autocorrelations autocorr MGLI-Langevin MGLI-Prior LI-Langevin LI-Prior H-Langevin PCN-RW autocorr MGLI-Langevin H-Langevin PCN-RW lag SNR lag SNR50

42 DILI samplers Sampling efficiency, conditioned diffusion (n=1000) DI and LI both helpful (c) autocorrelations autocorr MGLI-Langevin MGLI-Prior LI-Langevin LI-Prior H-Langevin PCN-RW lag

43 DILI samplers Sampling efficiency, conditioned diffusion (n=1000) Posterior-averaged Hessian yields improvements over MAP Hessian autocorr MAP-LIS Adapt-LIS lag

44 Ongoing work Transport-accelerated MCMC in high dimensions Combine transport maps with LIS ideas: map departure from identity limited to few directions Maps without MCMC: adaptive methods for high dimensions Maps from reference to target (rather than reverse) Enrichment via first variation of function-space optimization problem + progressive deflation Maps for nonlinear filtering Code (MUQ) available for download at: mituq/muq

45 References Moselhy, T; Marzouk YM. Bayesian inference with optimal maps. J. Comp. Phys. 281: (2012). Cui, T; Martin, J; Marzouk, YM; Solonen, A; Spantini, A. Likelihood-informed dimension reduction for nonlinear inverse probems. Inverse Problems 30: (2014). Spantini, A; Solonen, A; Cui, T; Martin, J; Tenorio, L; Marzouk, YM. Optimal lowrank approximation of linear Bayesian inverse problems. arxiv: (2014). Cui, T; Law, KJH; Marzouk, YM. Dimension independent likelihood informed MCMC. arxiv: (2014). Parno, M; Marzouk, YM. Transport map accelerated MCMC. arxiv: (2014).

46 Nonlinear Bayesian IP autocorrelation log-likelihood MCMC iteration 1st LIS vector rd LIS vector th LIS vector autocorrelation subspace full space CPU time (second)

Dimension-Independent likelihood-informed (DILI) MCMC

Dimension-Independent likelihood-informed (DILI) MCMC Dimension-Independent likelihood-informed (DILI) MCMC Tiangang Cui, Kody Law 2, Youssef Marzouk Massachusetts Institute of Technology 2 Oak Ridge National Laboratory 2 August 25 TC, KL, YM DILI MCMC USC

More information

The Bayesian approach to inverse problems

The Bayesian approach to inverse problems The Bayesian approach to inverse problems Youssef Marzouk Department of Aeronautics and Astronautics Center for Computational Engineering Massachusetts Institute of Technology ymarz@mit.edu, http://uqgroup.mit.edu

More information

Markov chain Monte Carlo methods

Markov chain Monte Carlo methods Markov chain Monte Carlo methods Youssef Marzouk Department of Aeronautics and Astronatics Massachusetts Institute of Technology ymarz@mit.edu 22 June 2015 Marzouk (MIT) IMA Summer School 22 June 2015

More information

Bayesian Methods and Uncertainty Quantification for Nonlinear Inverse Problems

Bayesian Methods and Uncertainty Quantification for Nonlinear Inverse Problems Bayesian Methods and Uncertainty Quantification for Nonlinear Inverse Problems John Bardsley, University of Montana Collaborators: H. Haario, J. Kaipio, M. Laine, Y. Marzouk, A. Seppänen, A. Solonen, Z.

More information

Monte Carlo in Bayesian Statistics

Monte Carlo in Bayesian Statistics Monte Carlo in Bayesian Statistics Matthew Thomas SAMBa - University of Bath m.l.thomas@bath.ac.uk December 4, 2014 Matthew Thomas (SAMBa) Monte Carlo in Bayesian Statistics December 4, 2014 1 / 16 Overview

More information

Bayesian Estimation of DSGE Models 1 Chapter 3: A Crash Course in Bayesian Inference

Bayesian Estimation of DSGE Models 1 Chapter 3: A Crash Course in Bayesian Inference 1 The views expressed in this paper are those of the authors and do not necessarily reflect the views of the Federal Reserve Board of Governors or the Federal Reserve System. Bayesian Estimation of DSGE

More information

MCMC Sampling for Bayesian Inference using L1-type Priors

MCMC Sampling for Bayesian Inference using L1-type Priors MÜNSTER MCMC Sampling for Bayesian Inference using L1-type Priors (what I do whenever the ill-posedness of EEG/MEG is just not frustrating enough!) AG Imaging Seminar Felix Lucka 26.06.2012 , MÜNSTER Sampling

More information

c 2016 Society for Industrial and Applied Mathematics

c 2016 Society for Industrial and Applied Mathematics SIAM J. SCI. COMPUT. Vol. 8, No. 5, pp. A779 A85 c 6 Society for Industrial and Applied Mathematics ACCELERATING MARKOV CHAIN MONTE CARLO WITH ACTIVE SUBSPACES PAUL G. CONSTANTINE, CARSON KENT, AND TAN

More information

Bayesian Inverse Problems with L[subscript 1] Priors: A Randomize-Then-Optimize Approach

Bayesian Inverse Problems with L[subscript 1] Priors: A Randomize-Then-Optimize Approach Bayesian Inverse Problems with L[subscript ] Priors: A Randomize-Then-Optimize Approach The MIT Faculty has made this article openly available. Please share how this access benefits you. Your story matters.

More information

Computer Practical: Metropolis-Hastings-based MCMC

Computer Practical: Metropolis-Hastings-based MCMC Computer Practical: Metropolis-Hastings-based MCMC Andrea Arnold and Franz Hamilton North Carolina State University July 30, 2016 A. Arnold / F. Hamilton (NCSU) MH-based MCMC July 30, 2016 1 / 19 Markov

More information

Manifold Monte Carlo Methods

Manifold Monte Carlo Methods Manifold Monte Carlo Methods Mark Girolami Department of Statistical Science University College London Joint work with Ben Calderhead Research Section Ordinary Meeting The Royal Statistical Society October

More information

LECTURE 15 Markov chain Monte Carlo

LECTURE 15 Markov chain Monte Carlo LECTURE 15 Markov chain Monte Carlo There are many settings when posterior computation is a challenge in that one does not have a closed form expression for the posterior distribution. Markov chain Monte

More information

Reminder of some Markov Chain properties:

Reminder of some Markov Chain properties: Reminder of some Markov Chain properties: 1. a transition from one state to another occurs probabilistically 2. only state that matters is where you currently are (i.e. given present, future is independent

More information

The University of Auckland Applied Mathematics Bayesian Methods for Inverse Problems : why and how Colin Fox Tiangang Cui, Mike O Sullivan (Auckland),

The University of Auckland Applied Mathematics Bayesian Methods for Inverse Problems : why and how Colin Fox Tiangang Cui, Mike O Sullivan (Auckland), The University of Auckland Applied Mathematics Bayesian Methods for Inverse Problems : why and how Colin Fox Tiangang Cui, Mike O Sullivan (Auckland), Geoff Nicholls (Statistics, Oxford) fox@math.auckland.ac.nz

More information

Adaptive Monte Carlo methods

Adaptive Monte Carlo methods Adaptive Monte Carlo methods Jean-Michel Marin Projet Select, INRIA Futurs, Université Paris-Sud joint with Randal Douc (École Polytechnique), Arnaud Guillin (Université de Marseille) and Christian Robert

More information

A Review of Pseudo-Marginal Markov Chain Monte Carlo

A Review of Pseudo-Marginal Markov Chain Monte Carlo A Review of Pseudo-Marginal Markov Chain Monte Carlo Discussed by: Yizhe Zhang October 21, 2016 Outline 1 Overview 2 Paper review 3 experiment 4 conclusion Motivation & overview Notation: θ denotes the

More information

Controlled sequential Monte Carlo

Controlled sequential Monte Carlo Controlled sequential Monte Carlo Jeremy Heng, Department of Statistics, Harvard University Joint work with Adrian Bishop (UTS, CSIRO), George Deligiannidis & Arnaud Doucet (Oxford) Bayesian Computation

More information

Riemann Manifold Methods in Bayesian Statistics

Riemann Manifold Methods in Bayesian Statistics Ricardo Ehlers ehlers@icmc.usp.br Applied Maths and Stats University of São Paulo, Brazil Working Group in Statistical Learning University College Dublin September 2015 Bayesian inference is based on Bayes

More information

Adaptive Posterior Approximation within MCMC

Adaptive Posterior Approximation within MCMC Adaptive Posterior Approximation within MCMC Tiangang Cui (MIT) Colin Fox (University of Otago) Mike O Sullivan (University of Auckland) Youssef Marzouk (MIT) Karen Willcox (MIT) 06/January/2012 C, F,

More information

Lecture 2: From Linear Regression to Kalman Filter and Beyond

Lecture 2: From Linear Regression to Kalman Filter and Beyond Lecture 2: From Linear Regression to Kalman Filter and Beyond January 18, 2017 Contents 1 Batch and Recursive Estimation 2 Towards Bayesian Filtering 3 Kalman Filter and Bayesian Filtering and Smoothing

More information

Kernel Sequential Monte Carlo

Kernel Sequential Monte Carlo Kernel Sequential Monte Carlo Ingmar Schuster (Paris Dauphine) Heiko Strathmann (University College London) Brooks Paige (Oxford) Dino Sejdinovic (Oxford) * equal contribution April 25, 2016 1 / 37 Section

More information

A Stein variational Newton method

A Stein variational Newton method A Stein variational Newton method Gianluca Detommaso University of Bath & The Alan Turing Institute gd39@bath.ac.uk Tiangang Cui Monash University Tiangang.Cui@monash.edu Alessio Spantini Massachusetts

More information

Risk Estimation and Uncertainty Quantification by Markov Chain Monte Carlo Methods

Risk Estimation and Uncertainty Quantification by Markov Chain Monte Carlo Methods Risk Estimation and Uncertainty Quantification by Markov Chain Monte Carlo Methods Konstantin Zuev Institute for Risk and Uncertainty University of Liverpool http://www.liv.ac.uk/risk-and-uncertainty/staff/k-zuev/

More information

Lecture 8: Bayesian Estimation of Parameters in State Space Models

Lecture 8: Bayesian Estimation of Parameters in State Space Models in State Space Models March 30, 2016 Contents 1 Bayesian estimation of parameters in state space models 2 Computational methods for parameter estimation 3 Practical parameter estimation in state space

More information

Robust MCMC Sampling with Non-Gaussian and Hierarchical Priors

Robust MCMC Sampling with Non-Gaussian and Hierarchical Priors Division of Engineering & Applied Science Robust MCMC Sampling with Non-Gaussian and Hierarchical Priors IPAM, UCLA, November 14, 2017 Matt Dunlop Victor Chen (Caltech) Omiros Papaspiliopoulos (ICREA,

More information

Learning the hyper-parameters. Luca Martino

Learning the hyper-parameters. Luca Martino Learning the hyper-parameters Luca Martino 2017 2017 1 / 28 Parameters and hyper-parameters 1. All the described methods depend on some choice of hyper-parameters... 2. For instance, do you recall λ (bandwidth

More information

Introduction to Machine Learning

Introduction to Machine Learning Introduction to Machine Learning Brown University CSCI 1950-F, Spring 2012 Prof. Erik Sudderth Lecture 25: Markov Chain Monte Carlo (MCMC) Course Review and Advanced Topics Many figures courtesy Kevin

More information

Kernel adaptive Sequential Monte Carlo

Kernel adaptive Sequential Monte Carlo Kernel adaptive Sequential Monte Carlo Ingmar Schuster (Paris Dauphine) Heiko Strathmann (University College London) Brooks Paige (Oxford) Dino Sejdinovic (Oxford) December 7, 2015 1 / 36 Section 1 Outline

More information

Gradient-based Monte Carlo sampling methods

Gradient-based Monte Carlo sampling methods Gradient-based Monte Carlo sampling methods Johannes von Lindheim 31. May 016 Abstract Notes for a 90-minute presentation on gradient-based Monte Carlo sampling methods for the Uncertainty Quantification

More information

Bayesian Inverse problem, Data assimilation and Localization

Bayesian Inverse problem, Data assimilation and Localization Bayesian Inverse problem, Data assimilation and Localization Xin T Tong National University of Singapore ICIP, Singapore 2018 X.Tong Localization 1 / 37 Content What is Bayesian inverse problem? What is

More information

Density Estimation. Seungjin Choi

Density Estimation. Seungjin Choi Density Estimation Seungjin Choi Department of Computer Science and Engineering Pohang University of Science and Technology 77 Cheongam-ro, Nam-gu, Pohang 37673, Korea seungjin@postech.ac.kr http://mlg.postech.ac.kr/

More information

Afternoon Meeting on Bayesian Computation 2018 University of Reading

Afternoon Meeting on Bayesian Computation 2018 University of Reading Gabriele Abbati 1, Alessra Tosi 2, Seth Flaxman 3, Michael A Osborne 1 1 University of Oxford, 2 Mind Foundry Ltd, 3 Imperial College London Afternoon Meeting on Bayesian Computation 2018 University of

More information

Sequential Monte Carlo Samplers for Applications in High Dimensions

Sequential Monte Carlo Samplers for Applications in High Dimensions Sequential Monte Carlo Samplers for Applications in High Dimensions Alexandros Beskos National University of Singapore KAUST, 26th February 2014 Joint work with: Dan Crisan, Ajay Jasra, Nik Kantas, Alex

More information

BAYESIAN METHODS FOR VARIABLE SELECTION WITH APPLICATIONS TO HIGH-DIMENSIONAL DATA

BAYESIAN METHODS FOR VARIABLE SELECTION WITH APPLICATIONS TO HIGH-DIMENSIONAL DATA BAYESIAN METHODS FOR VARIABLE SELECTION WITH APPLICATIONS TO HIGH-DIMENSIONAL DATA Intro: Course Outline and Brief Intro to Marina Vannucci Rice University, USA PASI-CIMAT 04/28-30/2010 Marina Vannucci

More information

arxiv: v4 [stat.co] 4 May 2016

arxiv: v4 [stat.co] 4 May 2016 Hamiltonian Monte Carlo Acceleration using Surrogate Functions with Random Bases arxiv:56.5555v4 [stat.co] 4 May 26 Cheng Zhang Department of Mathematics University of California, Irvine Irvine, CA 92697

More information

Stochastic Spectral Approaches to Bayesian Inference

Stochastic Spectral Approaches to Bayesian Inference Stochastic Spectral Approaches to Bayesian Inference Prof. Nathan L. Gibson Department of Mathematics Applied Mathematics and Computation Seminar March 4, 2011 Prof. Gibson (OSU) Spectral Approaches to

More information

MCMC for big data. Geir Storvik. BigInsight lunch - May Geir Storvik MCMC for big data BigInsight lunch - May / 17

MCMC for big data. Geir Storvik. BigInsight lunch - May Geir Storvik MCMC for big data BigInsight lunch - May / 17 MCMC for big data Geir Storvik BigInsight lunch - May 2 2018 Geir Storvik MCMC for big data BigInsight lunch - May 2 2018 1 / 17 Outline Why ordinary MCMC is not scalable Different approaches for making

More information

Physics 403. Segev BenZvi. Parameter Estimation, Correlations, and Error Bars. Department of Physics and Astronomy University of Rochester

Physics 403. Segev BenZvi. Parameter Estimation, Correlations, and Error Bars. Department of Physics and Astronomy University of Rochester Physics 403 Parameter Estimation, Correlations, and Error Bars Segev BenZvi Department of Physics and Astronomy University of Rochester Table of Contents 1 Review of Last Class Best Estimates and Reliability

More information

On Bayesian Computation

On Bayesian Computation On Bayesian Computation Michael I. Jordan with Elaine Angelino, Maxim Rabinovich, Martin Wainwright and Yun Yang Previous Work: Information Constraints on Inference Minimize the minimax risk under constraints

More information

17 : Optimization and Monte Carlo Methods

17 : Optimization and Monte Carlo Methods 10-708: Probabilistic Graphical Models Spring 2017 17 : Optimization and Monte Carlo Methods Lecturer: Avinava Dubey Scribes: Neil Spencer, YJ Choe 1 Recap 1.1 Monte Carlo Monte Carlo methods such as rejection

More information

The Particle Filter. PD Dr. Rudolph Triebel Computer Vision Group. Machine Learning for Computer Vision

The Particle Filter. PD Dr. Rudolph Triebel Computer Vision Group. Machine Learning for Computer Vision The Particle Filter Non-parametric implementation of Bayes filter Represents the belief (posterior) random state samples. by a set of This representation is approximate. Can represent distributions that

More information

MCMC algorithms for fitting Bayesian models

MCMC algorithms for fitting Bayesian models MCMC algorithms for fitting Bayesian models p. 1/1 MCMC algorithms for fitting Bayesian models Sudipto Banerjee sudiptob@biostat.umn.edu University of Minnesota MCMC algorithms for fitting Bayesian models

More information

Scalable algorithms for optimal experimental design for infinite-dimensional nonlinear Bayesian inverse problems

Scalable algorithms for optimal experimental design for infinite-dimensional nonlinear Bayesian inverse problems Scalable algorithms for optimal experimental design for infinite-dimensional nonlinear Bayesian inverse problems Alen Alexanderian (Math/NC State), Omar Ghattas (ICES/UT-Austin), Noémi Petra (Applied Math/UC

More information

A variational radial basis function approximation for diffusion processes

A variational radial basis function approximation for diffusion processes A variational radial basis function approximation for diffusion processes Michail D. Vrettas, Dan Cornford and Yuan Shen Aston University - Neural Computing Research Group Aston Triangle, Birmingham B4

More information

Slice Sampling with Adaptive Multivariate Steps: The Shrinking-Rank Method

Slice Sampling with Adaptive Multivariate Steps: The Shrinking-Rank Method Slice Sampling with Adaptive Multivariate Steps: The Shrinking-Rank Method Madeleine B. Thompson Radford M. Neal Abstract The shrinking rank method is a variation of slice sampling that is efficient at

More information

Advanced uncertainty evaluation of climate models by Monte Carlo methods

Advanced uncertainty evaluation of climate models by Monte Carlo methods Advanced uncertainty evaluation of climate models by Monte Carlo methods Marko Laine marko.laine@fmi.fi Pirkka Ollinaho, Janne Hakkarainen, Johanna Tamminen, Heikki Järvinen (FMI) Antti Solonen, Heikki

More information

arxiv: v2 [stat.ml] 29 Oct 2018

arxiv: v2 [stat.ml] 29 Oct 2018 A Stein variational Newton method Gianluca Detommaso University of Bath & The Alan Turing Institute gd39@bath.ac.uk Tiangang Cui Monash University Tiangang.Cui@monash.edu arxiv:86.385v [stat.ml] 9 Oct

More information

Bayesian Inference for Discretely Sampled Diffusion Processes: A New MCMC Based Approach to Inference

Bayesian Inference for Discretely Sampled Diffusion Processes: A New MCMC Based Approach to Inference Bayesian Inference for Discretely Sampled Diffusion Processes: A New MCMC Based Approach to Inference Osnat Stramer 1 and Matthew Bognar 1 Department of Statistics and Actuarial Science, University of

More information

PATTERN RECOGNITION AND MACHINE LEARNING

PATTERN RECOGNITION AND MACHINE LEARNING PATTERN RECOGNITION AND MACHINE LEARNING Chapter 1. Introduction Shuai Huang April 21, 2014 Outline 1 What is Machine Learning? 2 Curve Fitting 3 Probability Theory 4 Model Selection 5 The curse of dimensionality

More information

Integrated Non-Factorized Variational Inference

Integrated Non-Factorized Variational Inference Integrated Non-Factorized Variational Inference Shaobo Han, Xuejun Liao and Lawrence Carin Duke University February 27, 2014 S. Han et al. Integrated Non-Factorized Variational Inference February 27, 2014

More information

ICES REPORT March Tan Bui-Thanh And Mark Andrew Girolami

ICES REPORT March Tan Bui-Thanh And Mark Andrew Girolami ICES REPORT 4- March 4 Solving Large-Scale Pde-Constrained Bayesian Inverse Problems With Riemann Manifold Hamiltonian Monte Carlo by Tan Bui-Thanh And Mark Andrew Girolami The Institute for Computational

More information

The Metropolis-Hastings Algorithm. June 8, 2012

The Metropolis-Hastings Algorithm. June 8, 2012 The Metropolis-Hastings Algorithm June 8, 22 The Plan. Understand what a simulated distribution is 2. Understand why the Metropolis-Hastings algorithm works 3. Learn how to apply the Metropolis-Hastings

More information

Algorithms for Variational Learning of Mixture of Gaussians

Algorithms for Variational Learning of Mixture of Gaussians Algorithms for Variational Learning of Mixture of Gaussians Instructors: Tapani Raiko and Antti Honkela Bayes Group Adaptive Informatics Research Center 28.08.2008 Variational Bayesian Inference Mixture

More information

Bayesian Calibration of Simulators with Structured Discretization Uncertainty

Bayesian Calibration of Simulators with Structured Discretization Uncertainty Bayesian Calibration of Simulators with Structured Discretization Uncertainty Oksana A. Chkrebtii Department of Statistics, The Ohio State University Joint work with Matthew T. Pratola (Statistics, The

More information

arxiv: v1 [stat.co] 23 Apr 2018

arxiv: v1 [stat.co] 23 Apr 2018 Bayesian Updating and Uncertainty Quantification using Sequential Tempered MCMC with the Rank-One Modified Metropolis Algorithm Thomas A. Catanach and James L. Beck arxiv:1804.08738v1 [stat.co] 23 Apr

More information

Stochastic Collocation Methods for Polynomial Chaos: Analysis and Applications

Stochastic Collocation Methods for Polynomial Chaos: Analysis and Applications Stochastic Collocation Methods for Polynomial Chaos: Analysis and Applications Dongbin Xiu Department of Mathematics, Purdue University Support: AFOSR FA955-8-1-353 (Computational Math) SF CAREER DMS-64535

More information

Bayesian Inference and MCMC

Bayesian Inference and MCMC Bayesian Inference and MCMC Aryan Arbabi Partly based on MCMC slides from CSC412 Fall 2018 1 / 18 Bayesian Inference - Motivation Consider we have a data set D = {x 1,..., x n }. E.g each x i can be the

More information

Point spread function reconstruction from the image of a sharp edge

Point spread function reconstruction from the image of a sharp edge DOE/NV/5946--49 Point spread function reconstruction from the image of a sharp edge John Bardsley, Kevin Joyce, Aaron Luttman The University of Montana National Security Technologies LLC Montana Uncertainty

More information

Computational statistics

Computational statistics Computational statistics Markov Chain Monte Carlo methods Thierry Denœux March 2017 Thierry Denœux Computational statistics March 2017 1 / 71 Contents of this chapter When a target density f can be evaluated

More information

I. Bayesian econometrics

I. Bayesian econometrics I. Bayesian econometrics A. Introduction B. Bayesian inference in the univariate regression model C. Statistical decision theory D. Large sample results E. Diffuse priors F. Numerical Bayesian methods

More information

Expectation Propagation for Approximate Bayesian Inference

Expectation Propagation for Approximate Bayesian Inference Expectation Propagation for Approximate Bayesian Inference José Miguel Hernández Lobato Universidad Autónoma de Madrid, Computer Science Department February 5, 2007 1/ 24 Bayesian Inference Inference Given

More information

Variational Methods in Bayesian Deconvolution

Variational Methods in Bayesian Deconvolution PHYSTAT, SLAC, Stanford, California, September 8-, Variational Methods in Bayesian Deconvolution K. Zarb Adami Cavendish Laboratory, University of Cambridge, UK This paper gives an introduction to the

More information

Deblurring Jupiter (sampling in GLIP faster than regularized inversion) Colin Fox Richard A. Norton, J.

Deblurring Jupiter (sampling in GLIP faster than regularized inversion) Colin Fox Richard A. Norton, J. Deblurring Jupiter (sampling in GLIP faster than regularized inversion) Colin Fox fox@physics.otago.ac.nz Richard A. Norton, J. Andrés Christen Topics... Backstory (?) Sampling in linear-gaussian hierarchical

More information

Lecture 2: From Linear Regression to Kalman Filter and Beyond

Lecture 2: From Linear Regression to Kalman Filter and Beyond Lecture 2: From Linear Regression to Kalman Filter and Beyond Department of Biomedical Engineering and Computational Science Aalto University January 26, 2012 Contents 1 Batch and Recursive Estimation

More information

Nonparametric Bayesian Methods (Gaussian Processes)

Nonparametric Bayesian Methods (Gaussian Processes) [70240413 Statistical Machine Learning, Spring, 2015] Nonparametric Bayesian Methods (Gaussian Processes) Jun Zhu dcszj@mail.tsinghua.edu.cn http://bigml.cs.tsinghua.edu.cn/~jun State Key Lab of Intelligent

More information

Markov Chain Monte Carlo methods

Markov Chain Monte Carlo methods Markov Chain Monte Carlo methods Tomas McKelvey and Lennart Svensson Signal Processing Group Department of Signals and Systems Chalmers University of Technology, Sweden November 26, 2012 Today s learning

More information

Principles of Bayesian Inference

Principles of Bayesian Inference Principles of Bayesian Inference Sudipto Banerjee University of Minnesota July 20th, 2008 1 Bayesian Principles Classical statistics: model parameters are fixed and unknown. A Bayesian thinks of parameters

More information

Some Results on the Ergodicity of Adaptive MCMC Algorithms

Some Results on the Ergodicity of Adaptive MCMC Algorithms Some Results on the Ergodicity of Adaptive MCMC Algorithms Omar Khalil Supervisor: Jeffrey Rosenthal September 2, 2011 1 Contents 1 Andrieu-Moulines 4 2 Roberts-Rosenthal 7 3 Atchadé and Fort 8 4 Relationship

More information

Control Variates for Markov Chain Monte Carlo

Control Variates for Markov Chain Monte Carlo Control Variates for Markov Chain Monte Carlo Dellaportas, P., Kontoyiannis, I., and Tsourti, Z. Dept of Statistics, AUEB Dept of Informatics, AUEB 1st Greek Stochastics Meeting Monte Carlo: Probability

More information

1 Geometry of high dimensional probability distributions

1 Geometry of high dimensional probability distributions Hamiltonian Monte Carlo October 20, 2018 Debdeep Pati References: Neal, Radford M. MCMC using Hamiltonian dynamics. Handbook of Markov Chain Monte Carlo 2.11 (2011): 2. Betancourt, Michael. A conceptual

More information

Answers and expectations

Answers and expectations Answers and expectations For a function f(x) and distribution P(x), the expectation of f with respect to P is The expectation is the average of f, when x is drawn from the probability distribution P E

More information

Adaptive HMC via the Infinite Exponential Family

Adaptive HMC via the Infinite Exponential Family Adaptive HMC via the Infinite Exponential Family Arthur Gretton Gatsby Unit, CSML, University College London RegML, 2017 Arthur Gretton (Gatsby Unit, UCL) Adaptive HMC via the Infinite Exponential Family

More information

Stat 535 C - Statistical Computing & Monte Carlo Methods. Lecture February Arnaud Doucet

Stat 535 C - Statistical Computing & Monte Carlo Methods. Lecture February Arnaud Doucet Stat 535 C - Statistical Computing & Monte Carlo Methods Lecture 13-28 February 2006 Arnaud Doucet Email: arnaud@cs.ubc.ca 1 1.1 Outline Limitations of Gibbs sampling. Metropolis-Hastings algorithm. Proof

More information

Foundations of Statistical Inference

Foundations of Statistical Inference Foundations of Statistical Inference Julien Berestycki Department of Statistics University of Oxford MT 2016 Julien Berestycki (University of Oxford) SB2a MT 2016 1 / 32 Lecture 14 : Variational Bayes

More information

Likelihood-free MCMC

Likelihood-free MCMC Bayesian inference for stable distributions with applications in finance Department of Mathematics University of Leicester September 2, 2011 MSc project final presentation Outline 1 2 3 4 Classical Monte

More information

Transport maps for accelerated Bayesian computation. Matthew David Parno

Transport maps for accelerated Bayesian computation. Matthew David Parno Transport maps for accelerated Bayesian computation by Matthew David Parno S.M. Computation for Design and Optimization, Massachusetts Institute of Technology (2011) B.S. Applied Mathematics and Statistics,

More information

Bayesian parameter estimation in predictive engineering

Bayesian parameter estimation in predictive engineering Bayesian parameter estimation in predictive engineering Damon McDougall Institute for Computational Engineering and Sciences, UT Austin 14th August 2014 1/27 Motivation Understand physical phenomena Observations

More information

Exercises Tutorial at ICASSP 2016 Learning Nonlinear Dynamical Models Using Particle Filters

Exercises Tutorial at ICASSP 2016 Learning Nonlinear Dynamical Models Using Particle Filters Exercises Tutorial at ICASSP 216 Learning Nonlinear Dynamical Models Using Particle Filters Andreas Svensson, Johan Dahlin and Thomas B. Schön March 18, 216 Good luck! 1 [Bootstrap particle filter for

More information

Bayesian Prediction of Code Output. ASA Albuquerque Chapter Short Course October 2014

Bayesian Prediction of Code Output. ASA Albuquerque Chapter Short Course October 2014 Bayesian Prediction of Code Output ASA Albuquerque Chapter Short Course October 2014 Abstract This presentation summarizes Bayesian prediction methodology for the Gaussian process (GP) surrogate representation

More information

Multimodal Nested Sampling

Multimodal Nested Sampling Multimodal Nested Sampling Farhan Feroz Astrophysics Group, Cavendish Lab, Cambridge Inverse Problems & Cosmology Most obvious example: standard CMB data analysis pipeline But many others: object detection,

More information

STA 4273H: Sta-s-cal Machine Learning

STA 4273H: Sta-s-cal Machine Learning STA 4273H: Sta-s-cal Machine Learning Russ Salakhutdinov Department of Computer Science! Department of Statistical Sciences! rsalakhu@cs.toronto.edu! h0p://www.cs.utoronto.ca/~rsalakhu/ Lecture 2 In our

More information

Lecture 8: The Metropolis-Hastings Algorithm

Lecture 8: The Metropolis-Hastings Algorithm 30.10.2008 What we have seen last time: Gibbs sampler Key idea: Generate a Markov chain by updating the component of (X 1,..., X p ) in turn by drawing from the full conditionals: X (t) j Two drawbacks:

More information

IEOR E4570: Machine Learning for OR&FE Spring 2015 c 2015 by Martin Haugh. The EM Algorithm

IEOR E4570: Machine Learning for OR&FE Spring 2015 c 2015 by Martin Haugh. The EM Algorithm IEOR E4570: Machine Learning for OR&FE Spring 205 c 205 by Martin Haugh The EM Algorithm The EM algorithm is used for obtaining maximum likelihood estimates of parameters when some of the data is missing.

More information

Pattern Recognition and Machine Learning. Bishop Chapter 11: Sampling Methods

Pattern Recognition and Machine Learning. Bishop Chapter 11: Sampling Methods Pattern Recognition and Machine Learning Chapter 11: Sampling Methods Elise Arnaud Jakob Verbeek May 22, 2008 Outline of the chapter 11.1 Basic Sampling Algorithms 11.2 Markov Chain Monte Carlo 11.3 Gibbs

More information

Annealing Between Distributions by Averaging Moments

Annealing Between Distributions by Averaging Moments Annealing Between Distributions by Averaging Moments Chris J. Maddison Dept. of Comp. Sci. University of Toronto Roger Grosse CSAIL MIT Ruslan Salakhutdinov University of Toronto Partition Functions We

More information

Approximate Bayesian Computation: a simulation based approach to inference

Approximate Bayesian Computation: a simulation based approach to inference Approximate Bayesian Computation: a simulation based approach to inference Richard Wilkinson Simon Tavaré 2 Department of Probability and Statistics University of Sheffield 2 Department of Applied Mathematics

More information

Lecture 6: Bayesian Inference in SDE Models

Lecture 6: Bayesian Inference in SDE Models Lecture 6: Bayesian Inference in SDE Models Bayesian Filtering and Smoothing Point of View Simo Särkkä Aalto University Simo Särkkä (Aalto) Lecture 6: Bayesian Inference in SDEs 1 / 45 Contents 1 SDEs

More information

Introduction to Hamiltonian Monte Carlo Method

Introduction to Hamiltonian Monte Carlo Method Introduction to Hamiltonian Monte Carlo Method Mingwei Tang Department of Statistics University of Washington mingwt@uw.edu November 14, 2017 1 Hamiltonian System Notation: q R d : position vector, p R

More information

CPSC 540: Machine Learning

CPSC 540: Machine Learning CPSC 540: Machine Learning MCMC and Non-Parametric Bayes Mark Schmidt University of British Columbia Winter 2016 Admin I went through project proposals: Some of you got a message on Piazza. No news is

More information

Randomize-Then-Optimize for Sampling and Uncertainty Quantification in Electrical Impedance Tomography

Randomize-Then-Optimize for Sampling and Uncertainty Quantification in Electrical Impedance Tomography SIAM/ASA J. UNCERTAINTY QUANTIFICATION Vol. 3, pp. 1136 1158 c 2015 Society for Industrial and Applied Mathematics and American Statistical Association Randomize-Then-Optimize for Sampling and Uncertainty

More information

Markov chain Monte Carlo

Markov chain Monte Carlo 1 / 26 Markov chain Monte Carlo Timothy Hanson 1 and Alejandro Jara 2 1 Division of Biostatistics, University of Minnesota, USA 2 Department of Statistics, Universidad de Concepción, Chile IAP-Workshop

More information

MCMC and Gibbs Sampling. Kayhan Batmanghelich

MCMC and Gibbs Sampling. Kayhan Batmanghelich MCMC and Gibbs Sampling Kayhan Batmanghelich 1 Approaches to inference l Exact inference algorithms l l l The elimination algorithm Message-passing algorithm (sum-product, belief propagation) The junction

More information

An introduction to adaptive MCMC

An introduction to adaptive MCMC An introduction to adaptive MCMC Gareth Roberts MIRAW Day on Monte Carlo methods March 2011 Mainly joint work with Jeff Rosenthal. http://www2.warwick.ac.uk/fac/sci/statistics/crism/ Conferences and workshops

More information

Monte Carlo methods for sampling-based Stochastic Optimization

Monte Carlo methods for sampling-based Stochastic Optimization Monte Carlo methods for sampling-based Stochastic Optimization Gersende FORT LTCI CNRS & Telecom ParisTech Paris, France Joint works with B. Jourdain, T. Lelièvre, G. Stoltz from ENPC and E. Kuhn from

More information

ECE276A: Sensing & Estimation in Robotics Lecture 10: Gaussian Mixture and Particle Filtering

ECE276A: Sensing & Estimation in Robotics Lecture 10: Gaussian Mixture and Particle Filtering ECE276A: Sensing & Estimation in Robotics Lecture 10: Gaussian Mixture and Particle Filtering Lecturer: Nikolay Atanasov: natanasov@ucsd.edu Teaching Assistants: Siwei Guo: s9guo@eng.ucsd.edu Anwesan Pal:

More information

University of Toronto Department of Statistics

University of Toronto Department of Statistics Norm Comparisons for Data Augmentation by James P. Hobert Department of Statistics University of Florida and Jeffrey S. Rosenthal Department of Statistics University of Toronto Technical Report No. 0704

More information

Unsupervised Learning

Unsupervised Learning Unsupervised Learning Bayesian Model Comparison Zoubin Ghahramani zoubin@gatsby.ucl.ac.uk Gatsby Computational Neuroscience Unit, and MSc in Intelligent Systems, Dept Computer Science University College

More information

A SCALED STOCHASTIC NEWTON ALGORITHM FOR MARKOV CHAIN MONTE CARLO SIMULATIONS

A SCALED STOCHASTIC NEWTON ALGORITHM FOR MARKOV CHAIN MONTE CARLO SIMULATIONS A SCALED STOCHASTIC NEWTON ALGORITHM FOR MARKOV CHAIN MONTE CARLO SIMULATIONS TAN BUI-THANH AND OMAR GHATTAS Abstract. We propose a scaled stochastic Newton algorithm ssn) for local Metropolis-Hastings

More information

Efficient Variational Inference in Large-Scale Bayesian Compressed Sensing

Efficient Variational Inference in Large-Scale Bayesian Compressed Sensing Efficient Variational Inference in Large-Scale Bayesian Compressed Sensing George Papandreou and Alan Yuille Department of Statistics University of California, Los Angeles ICCV Workshop on Information

More information

19 : Slice Sampling and HMC

19 : Slice Sampling and HMC 10-708: Probabilistic Graphical Models 10-708, Spring 2018 19 : Slice Sampling and HMC Lecturer: Kayhan Batmanghelich Scribes: Boxiang Lyu 1 MCMC (Auxiliary Variables Methods) In inference, we are often

More information