Likelihood Inference for Lattice Spatial Processes

Size: px
Start display at page:

Download "Likelihood Inference for Lattice Spatial Processes"

Transcription

1 Likelihood Inference for Lattice Spatial Processes Donghoh Kim November 30, 2004 Donghoh Kim 1/24 Go to FULL

2 Lattice Processes Model : The Ising Model (1925), The Potts Model (1952), The Unitary Cell Model (1991) Phase Transition and Critical Value Likelihood Inference Simulation Donghoh Kim 2/24 Go to FULL

3 Lattice Processes Model : The Ising Model (1925), The Potts Model (1952), The Unitary Cell Model (1991) Phase Transition and Critical Value Likelihood Inference Simulation Donghoh Kim 2/24 Go to FULL

4 Lattice Spatial Processes Lattice i 1 i 2 i 3 i 4 i 5 i 6 i 7 i 8 i 9 β j jk k β jl l X i on each vertex takes q distinct values. neighbors : i 2, i 4, i 6, i 8 are nearest-neighbors of i 5 and i 1, i 3, i 5, i 7 are second-order-neighbors of i 5. Interaction : β ij between vertices. Lattice process models : unnormalized exponential family density on m n lattice. Donghoh Kim 3/24 Go to FULL

5 The Ising Model (Ising, 1925) A random variable X i takes two values in { 1, 1}. Only nearest-neighbors are considered. The one-parameter Ising model h(x) = exp(β t 2 (x)), β R. where t 2 (x) = i j x ix j, where means neighbor. t 2 (x) : no. of concordant pairs of variables minus no.of discordant pairs of variables. The two-parameter Ising model (Pickard, 1977) where h(x) = exp(β h t 2h (x) + β v t 2v (x)), t 2h (x) = i, j x ij x i(j+1), t 2v (x) = i, j x ij x (i+1)j Different strength of interaction between vertical nearest-neighbors and horizontal nearest-neighbors. Donghoh Kim 4/24 Go to FULL

6 2 2 Unitary Cell Model (Aguilar and Braun, 1991a) A 2 2 unitary cell : 2 2 small lattice with 8 interactions, repeated m n times. θ 11 θ 12 θ 11 θ 12 γ 11 θ 21 θ 22 θ 21 θ 22 γ 21 θ 11 θ 12 θ 11 θ 12 γ 11 γ 12 γ 12 γ 11 γ 11 γ 12 γ 12 θ 21 θ 22 θ 21 θ 22 γ 21 γ 22 γ 22 γ 21 γ 21 γ 22 γ X i takes only two values { 1, 1 }. h(x) = exp θ [i][j] x ij x i(j+1) + γ [i][j] x ij x (i+1)j i, j Includes the Ising model, lattice model, etc. by transforming the shape of the unitary cell and manipulating the interaction terms. Donghoh Kim 5/24 Go to FULL

7 Why Lattice Process Model? Spatial lattice data : plant ecology data (Besag 1974, Cressie 1993). Bayesian image restoration (Geman and Geman, 1984) : Lattice process prior Phase transitions. Ising model for magnetism Magnet consists of a large number of particles and each particle has two states N and S. At high temperature, loses magnetism and at low temperature, magnetized. Magnetisms occur suddenly (phase transitions) as the parameter value passes through a specific value, called the critical value (β = ). In statistical mechanics β = 1 kt, k; Boltzmann constant, T; Temperature. Donghoh Kim 6/24 Go to FULL

8 Phase Transitions and Critical Values The one-parameter Ising model (Onsager, 1944) : 0.5 sinh 1 (1) = The two-parameter Ising model (Pickard, 1977) : {(β h,β v ) : β h = 0.5tanh 1 (cos(ψ)), β v = 0.5tanh 1 (sin(ψ)),ψ (0, π/2)}. 2 2 Symmetric hexagonal lattice model (Aguilar and Braun, 1991b) : all six parameters are sinh 1 ( 3)/2 Donghoh Kim 7/24 Go to FULL

9 Likelihood Inference of Lattice Process Unormalized exponential family density : h θ (x), θ R k Normalizing constant : c(θ) = h θ (x) d(x) Normalized density : 1 c(θ) h θ(x) Problem c(θ) is unknown, so likelihood function is also unknown. Asymptotical distribution of MLE : Asymptotically calculate c(θ). Use Markov Chain Monte Carlo (MCMC) to approximate c(θ) (umbrella sampling). Donghoh Kim 8/24 Go to FULL

10 Asymptotic Distribution of MLE The one parameter Ising model When N = m n lattice size, β = 0.5 sinh 1 (1), ) ( 1 D (t 2(X N ) Et 2(X N ) N 0, 4 ). N log N π and ( N log N (ˆβ n β) D N 0, π ). 4 The two-parameter Ising model (Pickard, 1977) The critical surface : β h = 0.5 tanh 1 (cos(ψ)), β v = 0.5 tanh 1 (sin(ψ)) for ψ (0, π/2). where 1 N log N ( t2h (X N ) E t 2h (X N ) t 2v(X N ) E t 2v(X N ) Σ = 1 π ) ( tan ψ 1 1 cot ψ D N(0, Σ) ) Donghoh Kim 9/24 Go to FULL

11 Approximation of the Likelihood Function by MCMC A family {h θ : θ Θ} of unnormalized densities on state spacex. Normalizing constant c(θ) is unknown. The log likelihood : for an observation x, l(θ), ( ) 1 l(θ) = log c(θ) h θ(x) = log h θ (x) log c(θ). Take ratio against an arbitrary unnormalized density g. where c g = g(x)dx l(θ) = log h θ(x) g(x) log c(θ) c g, Domination Condition : Support of g contains support of h θ for all θ. c(θ) c g = E g h θ (X) g(x). Donghoh Kim 10/24 Go to FULL

12 Approximation of the Likelihood Function by MCMC Approximation : Generate an irreducible Markov chain X 1,...,X n from g, 1 n n i=1 h θ (X i ) g(x i ) a. s. c(θ). c g ( l(θ) l n (θ) = log h θ(x) g(x) log 1 n n i=1 Call g the background density and its sample the background sample. ) h θ (X i ). g(x i ) Maximization : Call MLE of approximated likelihood, Monte Carlo MLE (MCMLE). What distribution for g : any distribution as long as domination condition holds (cover large state space). Donghoh Kim 11/24 Go to FULL

13 Approximation of the Likelihood Function by MCMC Mixture Methods Simulated tempering (ST) (Marinari and Parisi,1992; Geyer and Thompson, 1995) Umbrella sampling (US) (Torrie and Valleau, 1977) To spread background samples out To approximate the likelihoods for a wide range of parameter values Move around finite set of distributions.... Donghoh Kim 12/24 Go to FULL

14 Approximation of the Likelihood Function by MCMC Mixture Methods The new state space pair (X, I), I indicating i-th unnormalized densities. Conditional on I = i, X has unnormalized density h ψi. Hence the joint distribution must be of the form h(x, i) = h ψi (x)a i where the a i are some constants called pseudo-prior. h(x, i) is used as background density. Donghoh Kim 13/24 Go to FULL

15 Approximation of the Likelihood Function by MCMC Mixture Methods One-parameter case (Geyer and Thompson, 1995) 1. Given the m distributions h ψ1,...,h ψm Update X h(x I n = i) = 1 c(ψ i ) h ψ i (x) Already have updating scheme for h ψi (x), Metropolis-Hastings or Gibbs. Update I using a Metropolis-Hastings. Proposal density q(i, j) for I, q(i, j) = q ij, j = i ± Add more distributions and repeat (1). Donghoh Kim 14/24 Go to FULL

16 Approximation of the Likelihood Function by MCMC Mixture Methods How to choose pseudo-prior a i The unnormalized marginal density of I is h(x, i)dx = c(ψ i )a i, c(ψ i ) = h ψi (x)dx. If pseudo-prior a i = 1/c(ψ i ), the marginal density of I is uniform. Any a i is working, but when uneven, sampler does not mix well. Geyer and Thompson (1995) adjust a i as a i /o i where o i is the occupation number of the i-th unnormalized density in an MCMC run. How to choose the m distributions Arrange m distributions according to parameter value and call spacing the distance between parameters. Adjusting spacing, and adding more distributions in an MCMC run so that the acceptance rates in Metropolis-Hastings updates for I are between 20% and 40%. Donghoh Kim 15/24 Go to FULL

17 Approximation of the Likelihood Function by MCMC Mixture Methods Multi-parameter case : Open Problem How to choose m distributions? There is no natural ordering of distributions How many Distributions? Need large number of distributions. Difficult to adjust the a i using the method of Geyer and Thompson (1995) Torrie and Valleau (1977) did not give any clear proposals. Donghoh Kim 16/24 Go to FULL

18 Umbrella sampling 1. Given mixture distribution h ψ1,..., h ψm in the model, and Given a i 2. Generate preliminary background samples Update state space X given i. h(x i) = h ψi (x) Update I given x by Gibbs update. h(i x) = h ψi (x) a i m k=1 h ψ k (x) a k. 3. Tuning pseudo-prior a k : Adjusting nc by mixture distribution itself and a k = 1/c(ψ k ) 1 n n i=1 h ψk (X Ik ) h ψii (X Ii )a Ii 4. Generate mixture background samples. 5. If satisfy, stop. Otherwise, add more distributions and repeat step 2. Donghoh Kim 17/24 Go to FULL

19 Umbrella sampling 2000 observations x from the symmetric hexagonal lattice model h θ. All six parameters are sinh 1 ( 3)/ Quantiles of Chi-Square Distribution with 6 d.f. Quantile plot of (t(x) Et(X)) ˆK 1 (t(x) Et(X)) If t(x) close to a normal, then the constrained MLE ˆθ close to N(θ, Σ) and Q close to χ 2 with 5 d.f., where Q = (ˆθ θ) Σ (ˆθ θ) Donghoh Kim 18/24 Go to FULL

20 Approximate likelihood by g. ( l(θ) l n (θ) = log h θ(x) g(x) log 1 n n i=1 ) h θ (X i ). g(x i ) MCMLE ˆθ satisfying constraint (critical equation) and where l n (θ) = t(x) n t(x i )w θ (X i ) = 0, i=1 w θ (x) = h θ(x) g(x) / n i=1 h θ (X i ) g(x i ). If an observation is on the boundary of the convex hull, the MCMLE does not exist. Donghoh Kim 19/24 Go to FULL

21 Umbrella sampling In the first run, 36 distributions are chosen up to 60 percentile of Q. MCMLE does not exist for 10 observations. In the second run, adding 20 more distributions between 60 and 80 percentile of Q. MCMLE dose not exist for 5 observations. In the third run, adding 45 more distributions from 80 to 95 percentile of Q. Why 60, 80, 95 percentile? not too far away from previous mixture : difficult to estimate the normalizing constant since the state space of new mixture is not covered. not too close to previous mixture :convex hull of the mixture is not spread out. As long as sampler mixes well. Donghoh Kim 20/24 Go to FULL

22 Quantiles of Chi-Square Distribution with 5 d.f. Quantile-Quantile plot of Q, Donghoh Kim 21/24 Go to FULL

23 Markov Chain Monte Carlo (MCMC) Use a Markov Chain X 0, X 1,... satisfying P(X n+1 X 0, X 1,...,X n ) = P(X n+1 X n ). Distribution of a Markov chain is defined through an initial distribution ν, i.e. the distribution of X 0 and transition probability P(, ). π is called invariant if π(a) = π(dx)p(x, A), for A S. Construct P(x, A) so that the Markov chain has π as its unique invariant distribution. Random vector X = (X 1, X 2,...,X k ) in State space S = S 1 S k with distribution π w.r.t. µ = µ 1 µ k. Let h be an unnormalized density w.r.t. the µ. Let X 1 be the variable to be updated and let X j = (X 1,...,X j 1, X j+1,...,x k ). Donghoh Kim 22/24 Go to FULL

24 Markov Chain Monte Carlo (MCMC) The Gibbs update : Update X 1 by π( X 1 ). The Metropolis-Hastings update : Current state x = (x 1,x 2,...,x k ) S. The proposal density q 1 (x, ) propose a candidate value y 1 S 1. Proposal y is (y 1,x 2,...,x k ). y is accepted with probability a(x 1,y 1 ), where a(x 1,y 1 ) = min(1,r(x 1,y 1 )), R(x 1,y 1 ) = h(y)q 1(y,x 1 ) h(x)q 1 (x,y 1 ) Hastings Ratio. The Metropolis-Hastings update may reject a new proposal whereas the Gibbs update always accepts a new proposal. Donghoh Kim 23/24 Go to FULL

Computer Vision Group Prof. Daniel Cremers. 11. Sampling Methods: Markov Chain Monte Carlo

Computer Vision Group Prof. Daniel Cremers. 11. Sampling Methods: Markov Chain Monte Carlo Group Prof. Daniel Cremers 11. Sampling Methods: Markov Chain Monte Carlo Markov Chain Monte Carlo In high-dimensional spaces, rejection sampling and importance sampling are very inefficient An alternative

More information

A = {(x, u) : 0 u f(x)},

A = {(x, u) : 0 u f(x)}, Draw x uniformly from the region {x : f(x) u }. Markov Chain Monte Carlo Lecture 5 Slice sampler: Suppose that one is interested in sampling from a density f(x), x X. Recall that sampling x f(x) is equivalent

More information

A Search and Jump Algorithm for Markov Chain Monte Carlo Sampling. Christopher Jennison. Adriana Ibrahim. Seminar at University of Kuwait

A Search and Jump Algorithm for Markov Chain Monte Carlo Sampling. Christopher Jennison. Adriana Ibrahim. Seminar at University of Kuwait A Search and Jump Algorithm for Markov Chain Monte Carlo Sampling Christopher Jennison Department of Mathematical Sciences, University of Bath, UK http://people.bath.ac.uk/mascj Adriana Ibrahim Institute

More information

Markov Chain Monte Carlo Inference. Siamak Ravanbakhsh Winter 2018

Markov Chain Monte Carlo Inference. Siamak Ravanbakhsh Winter 2018 Graphical Models Markov Chain Monte Carlo Inference Siamak Ravanbakhsh Winter 2018 Learning objectives Markov chains the idea behind Markov Chain Monte Carlo (MCMC) two important examples: Gibbs sampling

More information

Bayesian Inference in GLMs. Frequentists typically base inferences on MLEs, asymptotic confidence

Bayesian Inference in GLMs. Frequentists typically base inferences on MLEs, asymptotic confidence Bayesian Inference in GLMs Frequentists typically base inferences on MLEs, asymptotic confidence limits, and log-likelihood ratio tests Bayesians base inferences on the posterior distribution of the unknowns

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

Markov Chain Monte Carlo Lecture 4

Markov Chain Monte Carlo Lecture 4 The local-trap problem refers to that in simulations of a complex system whose energy landscape is rugged, the sampler gets trapped in a local energy minimum indefinitely, rendering the simulation ineffective.

More information

Introduction to Machine Learning CMU-10701

Introduction to Machine Learning CMU-10701 Introduction to Machine Learning CMU-10701 Markov Chain Monte Carlo Methods Barnabás Póczos & Aarti Singh Contents Markov Chain Monte Carlo Methods Goal & Motivation Sampling Rejection Importance Markov

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

i=1 h n (ˆθ n ) = 0. (2)

i=1 h n (ˆθ n ) = 0. (2) Stat 8112 Lecture Notes Unbiased Estimating Equations Charles J. Geyer April 29, 2012 1 Introduction In this handout we generalize the notion of maximum likelihood estimation to solution of unbiased estimating

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

Bayesian Methods for Machine Learning

Bayesian Methods for Machine Learning Bayesian Methods for Machine Learning CS 584: Big Data Analytics Material adapted from Radford Neal s tutorial (http://ftp.cs.utoronto.ca/pub/radford/bayes-tut.pdf), Zoubin Ghahramni (http://hunch.net/~coms-4771/zoubin_ghahramani_bayesian_learning.pdf),

More information

Monte Carlo Dynamically Weighted Importance Sampling for Spatial Models with Intractable Normalizing Constants

Monte Carlo Dynamically Weighted Importance Sampling for Spatial Models with Intractable Normalizing Constants Monte Carlo Dynamically Weighted Importance Sampling for Spatial Models with Intractable Normalizing Constants Faming Liang Texas A& University Sooyoung Cheon Korea University Spatial Model Introduction

More information

Markov Chain Monte Carlo (MCMC)

Markov Chain Monte Carlo (MCMC) Markov Chain Monte Carlo (MCMC Dependent Sampling Suppose we wish to sample from a density π, and we can evaluate π as a function but have no means to directly generate a sample. Rejection sampling can

More information

Computer Vision Group Prof. Daniel Cremers. 10a. Markov Chain Monte Carlo

Computer Vision Group Prof. Daniel Cremers. 10a. Markov Chain Monte Carlo Group Prof. Daniel Cremers 10a. Markov Chain Monte Carlo Markov Chain Monte Carlo In high-dimensional spaces, rejection sampling and importance sampling are very inefficient An alternative is Markov Chain

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

17 : Markov Chain Monte Carlo

17 : Markov Chain Monte Carlo 10-708: Probabilistic Graphical Models, Spring 2015 17 : Markov Chain Monte Carlo Lecturer: Eric P. Xing Scribes: Heran Lin, Bin Deng, Yun Huang 1 Review of Monte Carlo Methods 1.1 Overview Monte Carlo

More information

MCMC: Markov Chain Monte Carlo

MCMC: Markov Chain Monte Carlo I529: Machine Learning in Bioinformatics (Spring 2013) MCMC: Markov Chain Monte Carlo Yuzhen Ye School of Informatics and Computing Indiana University, Bloomington Spring 2013 Contents Review of Markov

More information

Statistics & Data Sciences: First Year Prelim Exam May 2018

Statistics & Data Sciences: First Year Prelim Exam May 2018 Statistics & Data Sciences: First Year Prelim Exam May 2018 Instructions: 1. Do not turn this page until instructed to do so. 2. Start each new question on a new sheet of paper. 3. This is a closed book

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

Bayesian Graphical Models

Bayesian Graphical Models Graphical Models and Inference, Lecture 16, Michaelmas Term 2009 December 4, 2009 Parameter θ, data X = x, likelihood L(θ x) p(x θ). Express knowledge about θ through prior distribution π on θ. Inference

More information

Convex Optimization CMU-10725

Convex Optimization CMU-10725 Convex Optimization CMU-10725 Simulated Annealing Barnabás Póczos & Ryan Tibshirani Andrey Markov Markov Chains 2 Markov Chains Markov chain: Homogen Markov chain: 3 Markov Chains Assume that the state

More information

Stat 8501 Lecture Notes Spatial Lattice Processes Charles J. Geyer November 16, Introduction

Stat 8501 Lecture Notes Spatial Lattice Processes Charles J. Geyer November 16, Introduction Stat 8501 Lecture Notes Spatial Lattice Processes Charles J. Geyer November 16, 2016 1 Introduction A spatial lattice process is a stochastic process with a discrete carrier T, that is, we have random

More information

Exponential families also behave nicely under conditioning. Specifically, suppose we write η = (η 1, η 2 ) R k R p k so that

Exponential families also behave nicely under conditioning. Specifically, suppose we write η = (η 1, η 2 ) R k R p k so that 1 More examples 1.1 Exponential families under conditioning Exponential families also behave nicely under conditioning. Specifically, suppose we write η = η 1, η 2 R k R p k so that dp η dm 0 = e ηt 1

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

Computer Vision Group Prof. Daniel Cremers. 14. Sampling Methods

Computer Vision Group Prof. Daniel Cremers. 14. Sampling Methods Prof. Daniel Cremers 14. Sampling Methods Sampling Methods Sampling Methods are widely used in Computer Science as an approximation of a deterministic algorithm to represent uncertainty without a parametric

More information

Markov Chain Monte Carlo

Markov Chain Monte Carlo 1 Motivation 1.1 Bayesian Learning Markov Chain Monte Carlo Yale Chang In Bayesian learning, given data X, we make assumptions on the generative process of X by introducing hidden variables Z: p(z): prior

More information

Master s Written Examination

Master s Written Examination Master s Written Examination Option: Statistics and Probability Spring 05 Full points may be obtained for correct answers to eight questions Each numbered question (which may have several parts) is worth

More information

Generalized Rejection Sampling Schemes and Applications in Signal Processing

Generalized Rejection Sampling Schemes and Applications in Signal Processing Generalized Rejection Sampling Schemes and Applications in Signal Processing 1 arxiv:0904.1300v1 [stat.co] 8 Apr 2009 Luca Martino and Joaquín Míguez Department of Signal Theory and Communications, Universidad

More information

SAMPLING ALGORITHMS. In general. Inference in Bayesian models

SAMPLING ALGORITHMS. In general. Inference in Bayesian models SAMPLING ALGORITHMS SAMPLING ALGORITHMS In general A sampling algorithm is an algorithm that outputs samples x 1, x 2,... from a given distribution P or density p. Sampling algorithms can for example be

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

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

MH I. Metropolis-Hastings (MH) algorithm is the most popular method of getting dependent samples from a probability distribution

MH I. Metropolis-Hastings (MH) algorithm is the most popular method of getting dependent samples from a probability distribution MH I Metropolis-Hastings (MH) algorithm is the most popular method of getting dependent samples from a probability distribution a lot of Bayesian mehods rely on the use of MH algorithm and it s famous

More information

3 : Representation of Undirected GM

3 : Representation of Undirected GM 10-708: Probabilistic Graphical Models 10-708, Spring 2016 3 : Representation of Undirected GM Lecturer: Eric P. Xing Scribes: Longqi Cai, Man-Chia Chang 1 MRF vs BN There are two types of graphical models:

More information

Learning MN Parameters with Approximation. Sargur Srihari

Learning MN Parameters with Approximation. Sargur Srihari Learning MN Parameters with Approximation Sargur srihari@cedar.buffalo.edu 1 Topics Iterative exact learning of MN parameters Difficulty with exact methods Approximate methods Approximate Inference Belief

More information

Monte Carlo Methods. Leon Gu CSD, CMU

Monte Carlo Methods. Leon Gu CSD, CMU Monte Carlo Methods Leon Gu CSD, CMU Approximate Inference EM: y-observed variables; x-hidden variables; θ-parameters; E-step: q(x) = p(x y, θ t 1 ) M-step: θ t = arg max E q(x) [log p(y, x θ)] θ Monte

More information

The Ising model and Markov chain Monte Carlo

The Ising model and Markov chain Monte Carlo The Ising model and Markov chain Monte Carlo Ramesh Sridharan These notes give a short description of the Ising model for images and an introduction to Metropolis-Hastings and Gibbs Markov Chain Monte

More information

Stat 516, Homework 1

Stat 516, Homework 1 Stat 516, Homework 1 Due date: October 7 1. Consider an urn with n distinct balls numbered 1,..., n. We sample balls from the urn with replacement. Let N be the number of draws until we encounter a ball

More information

(5) Multi-parameter models - Gibbs sampling. ST440/540: Applied Bayesian Analysis

(5) Multi-parameter models - Gibbs sampling. ST440/540: Applied Bayesian Analysis Summarizing a posterior Given the data and prior the posterior is determined Summarizing the posterior gives parameter estimates, intervals, and hypothesis tests Most of these computations are integrals

More information

CS242: Probabilistic Graphical Models Lecture 7B: Markov Chain Monte Carlo & Gibbs Sampling

CS242: Probabilistic Graphical Models Lecture 7B: Markov Chain Monte Carlo & Gibbs Sampling CS242: Probabilistic Graphical Models Lecture 7B: Markov Chain Monte Carlo & Gibbs Sampling Professor Erik Sudderth Brown University Computer Science October 27, 2016 Some figures and materials courtesy

More information

Nuisance parameters and their treatment

Nuisance parameters and their treatment BS2 Statistical Inference, Lecture 2, Hilary Term 2008 April 2, 2008 Ancillarity Inference principles Completeness A statistic A = a(x ) is said to be ancillary if (i) The distribution of A does not depend

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

6 Markov Chain Monte Carlo (MCMC)

6 Markov Chain Monte Carlo (MCMC) 6 Markov Chain Monte Carlo (MCMC) The underlying idea in MCMC is to replace the iid samples of basic MC methods, with dependent samples from an ergodic Markov chain, whose limiting (stationary) distribution

More information

Probabilistic Graphical Models

Probabilistic Graphical Models 2016 Robert Nowak Probabilistic Graphical Models 1 Introduction We have focused mainly on linear models for signals, in particular the subspace model x = Uθ, where U is a n k matrix and θ R k is a vector

More information

Markov Chain Monte Carlo

Markov Chain Monte Carlo Chapter 5 Markov Chain Monte Carlo MCMC is a kind of improvement of the Monte Carlo method By sampling from a Markov chain whose stationary distribution is the desired sampling distributuion, it is possible

More information

Theory of Stochastic Processes 8. Markov chain Monte Carlo

Theory of Stochastic Processes 8. Markov chain Monte Carlo Theory of Stochastic Processes 8. Markov chain Monte Carlo Tomonari Sei sei@mist.i.u-tokyo.ac.jp Department of Mathematical Informatics, University of Tokyo June 8, 2017 http://www.stat.t.u-tokyo.ac.jp/~sei/lec.html

More information

Generalized Exponential Random Graph Models: Inference for Weighted Graphs

Generalized Exponential Random Graph Models: Inference for Weighted Graphs Generalized Exponential Random Graph Models: Inference for Weighted Graphs James D. Wilson University of North Carolina at Chapel Hill June 18th, 2015 Political Networks, 2015 James D. Wilson GERGMs for

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

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

Computer Intensive Methods in Mathematical Statistics

Computer Intensive Methods in Mathematical Statistics Computer Intensive Methods in Mathematical Statistics Department of mathematics johawes@kth.se Lecture 16 Advanced topics in computational statistics 18 May 2017 Computer Intensive Methods (1) Plan of

More information

Introduction to Bayesian methods in inverse problems

Introduction to Bayesian methods in inverse problems Introduction to Bayesian methods in inverse problems Ville Kolehmainen 1 1 Department of Applied Physics, University of Eastern Finland, Kuopio, Finland March 4 2013 Manchester, UK. Contents Introduction

More information

Graphical Models and Kernel Methods

Graphical Models and Kernel Methods Graphical Models and Kernel Methods Jerry Zhu Department of Computer Sciences University of Wisconsin Madison, USA MLSS June 17, 2014 1 / 123 Outline Graphical Models Probabilistic Inference Directed vs.

More information

Introduction to Graphical Models

Introduction to Graphical Models Introduction to Graphical Models STA 345: Multivariate Analysis Department of Statistical Science Duke University, Durham, NC, USA Robert L. Wolpert 1 Conditional Dependence Two real-valued or vector-valued

More information

General Construction of Irreversible Kernel in Markov Chain Monte Carlo

General Construction of Irreversible Kernel in Markov Chain Monte Carlo General Construction of Irreversible Kernel in Markov Chain Monte Carlo Metropolis heat bath Suwa Todo Department of Applied Physics, The University of Tokyo Department of Physics, Boston University (from

More information

Down by the Bayes, where the Watermelons Grow

Down by the Bayes, where the Watermelons Grow Down by the Bayes, where the Watermelons Grow A Bayesian example using SAS SUAVe: Victoria SAS User Group Meeting November 21, 2017 Peter K. Ott, M.Sc., P.Stat. Strategic Analysis 1 Outline 1. Motivating

More information

Energy Based Models. Stefano Ermon, Aditya Grover. Stanford University. Lecture 13

Energy Based Models. Stefano Ermon, Aditya Grover. Stanford University. Lecture 13 Energy Based Models Stefano Ermon, Aditya Grover Stanford University Lecture 13 Stefano Ermon, Aditya Grover (AI Lab) Deep Generative Models Lecture 13 1 / 21 Summary Story so far Representation: Latent

More information

Brief introduction to Markov Chain Monte Carlo

Brief introduction to Markov Chain Monte Carlo Brief introduction to Department of Probability and Mathematical Statistics seminar Stochastic modeling in economics and finance November 7, 2011 Brief introduction to Content 1 and motivation Classical

More information

Probabilistic Graphical Models Lecture Notes Fall 2009

Probabilistic Graphical Models Lecture Notes Fall 2009 Probabilistic Graphical Models Lecture Notes Fall 2009 October 28, 2009 Byoung-Tak Zhang School of omputer Science and Engineering & ognitive Science, Brain Science, and Bioinformatics Seoul National University

More information

Computational Statistics. Jian Pei School of Computing Science Simon Fraser University

Computational Statistics. Jian Pei School of Computing Science Simon Fraser University Computational Statistics Jian Pei School of Computing Science Simon Fraser University jpei@cs.sfu.ca BASIC OPTIMIZATION METHODS J. Pei: Computational Statistics 2 Why Optimization? In statistical inference,

More information

Markov Chain Monte Carlo, Numerical Integration

Markov Chain Monte Carlo, Numerical Integration Markov Chain Monte Carlo, Numerical Integration (See Statistics) Trevor Gallen Fall 2015 1 / 1 Agenda Numerical Integration: MCMC methods Estimating Markov Chains Estimating latent variables 2 / 1 Numerical

More information

Basic math for biology

Basic math for biology Basic math for biology Lei Li Florida State University, Feb 6, 2002 The EM algorithm: setup Parametric models: {P θ }. Data: full data (Y, X); partial data Y. Missing data: X. Likelihood and maximum likelihood

More information

Patterns of Scalable Bayesian Inference Background (Session 1)

Patterns of Scalable Bayesian Inference Background (Session 1) Patterns of Scalable Bayesian Inference Background (Session 1) Jerónimo Arenas-García Universidad Carlos III de Madrid jeronimo.arenas@gmail.com June 14, 2017 1 / 15 Motivation. Bayesian Learning principles

More information

The Origin of Deep Learning. Lili Mou Jan, 2015

The Origin of Deep Learning. Lili Mou Jan, 2015 The Origin of Deep Learning Lili Mou Jan, 2015 Acknowledgment Most of the materials come from G. E. Hinton s online course. Outline Introduction Preliminary Boltzmann Machines and RBMs Deep Belief Nets

More information

Introduction to Restricted Boltzmann Machines

Introduction to Restricted Boltzmann Machines Introduction to Restricted Boltzmann Machines Ilija Bogunovic and Edo Collins EPFL {ilija.bogunovic,edo.collins}@epfl.ch October 13, 2014 Introduction Ingredients: 1. Probabilistic graphical models (undirected,

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

Normalising constants and maximum likelihood inference

Normalising constants and maximum likelihood inference Normalising constants and maximum likelihood inference Jakob G. Rasmussen Department of Mathematics Aalborg University Denmark March 9, 2011 1/14 Today Normalising constants Approximation of normalising

More information

ST 740: Markov Chain Monte Carlo

ST 740: Markov Chain Monte Carlo ST 740: Markov Chain Monte Carlo Alyson Wilson Department of Statistics North Carolina State University October 14, 2012 A. Wilson (NCSU Stsatistics) MCMC October 14, 2012 1 / 20 Convergence Diagnostics:

More information

Probabilistic Graphical Models Lecture 17: Markov chain Monte Carlo

Probabilistic Graphical Models Lecture 17: Markov chain Monte Carlo Probabilistic Graphical Models Lecture 17: Markov chain Monte Carlo Andrew Gordon Wilson www.cs.cmu.edu/~andrewgw Carnegie Mellon University March 18, 2015 1 / 45 Resources and Attribution Image credits,

More information

Markov Chain Monte Carlo Lecture 6

Markov Chain Monte Carlo Lecture 6 Sequential parallel tempering With the development of science and technology, we more and more need to deal with high dimensional systems. For example, we need to align a group of protein or DNA sequences

More information

Probabilistic Graphical Models

Probabilistic Graphical Models School of Computer Science Probabilistic Graphical Models Variational Inference II: Mean Field Method and Variational Principle Junming Yin Lecture 15, March 7, 2012 X 1 X 1 X 1 X 1 X 2 X 3 X 2 X 2 X 3

More information

Computer Vision Group Prof. Daniel Cremers. 11. Sampling Methods

Computer Vision Group Prof. Daniel Cremers. 11. Sampling Methods Prof. Daniel Cremers 11. Sampling Methods Sampling Methods Sampling Methods are widely used in Computer Science as an approximation of a deterministic algorithm to represent uncertainty without a parametric

More information

Markov Chain Monte Carlo Using the Ratio-of-Uniforms Transformation. Luke Tierney Department of Statistics & Actuarial Science University of Iowa

Markov Chain Monte Carlo Using the Ratio-of-Uniforms Transformation. Luke Tierney Department of Statistics & Actuarial Science University of Iowa Markov Chain Monte Carlo Using the Ratio-of-Uniforms Transformation Luke Tierney Department of Statistics & Actuarial Science University of Iowa Basic Ratio of Uniforms Method Introduced by Kinderman and

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

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

Simulation - Lectures - Part III Markov chain Monte Carlo

Simulation - Lectures - Part III Markov chain Monte Carlo Simulation - Lectures - Part III Markov chain Monte Carlo Julien Berestycki Part A Simulation and Statistical Programming Hilary Term 2018 Part A Simulation. HT 2018. J. Berestycki. 1 / 50 Outline Markov

More information

Markov Random Fields and Bayesian Image Analysis. Wei Liu Advisor: Tom Fletcher

Markov Random Fields and Bayesian Image Analysis. Wei Liu Advisor: Tom Fletcher Markov Random Fields and Bayesian Image Analysis Wei Liu Advisor: Tom Fletcher 1 Markov Random Field: Application Overview Awate and Whitaker 2006 2 Markov Random Field: Application Overview 3 Markov Random

More information

Kazuhiko Kakamu Department of Economics Finance, Institute for Advanced Studies. Abstract

Kazuhiko Kakamu Department of Economics Finance, Institute for Advanced Studies. Abstract Bayesian Estimation of A Distance Functional Weight Matrix Model Kazuhiko Kakamu Department of Economics Finance, Institute for Advanced Studies Abstract This paper considers the distance functional weight

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

The Pennsylvania State University The Graduate School RATIO-OF-UNIFORMS MARKOV CHAIN MONTE CARLO FOR GAUSSIAN PROCESS MODELS

The Pennsylvania State University The Graduate School RATIO-OF-UNIFORMS MARKOV CHAIN MONTE CARLO FOR GAUSSIAN PROCESS MODELS The Pennsylvania State University The Graduate School RATIO-OF-UNIFORMS MARKOV CHAIN MONTE CARLO FOR GAUSSIAN PROCESS MODELS A Thesis in Statistics by Chris Groendyke c 2008 Chris Groendyke Submitted in

More information

CSC 412 (Lecture 4): Undirected Graphical Models

CSC 412 (Lecture 4): Undirected Graphical Models CSC 412 (Lecture 4): Undirected Graphical Models Raquel Urtasun University of Toronto Feb 2, 2016 R Urtasun (UofT) CSC 412 Feb 2, 2016 1 / 37 Today Undirected Graphical Models: Semantics of the graph:

More information

Lecture 6: Graphical Models: Learning

Lecture 6: Graphical Models: Learning Lecture 6: Graphical Models: Learning 4F13: Machine Learning Zoubin Ghahramani and Carl Edward Rasmussen Department of Engineering, University of Cambridge February 3rd, 2010 Ghahramani & Rasmussen (CUED)

More information

Lecture 8: Computer Simulations of Generalized Ensembles

Lecture 8: Computer Simulations of Generalized Ensembles Lecture 8: Computer Simulations of Generalized Ensembles Bernd A. Berg Florida State University November 6, 2008 Bernd A. Berg (FSU) Generalized Ensembles November 6, 2008 1 / 33 Overview 1. Reweighting

More information

Parameter estimation and forecasting. Cristiano Porciani AIfA, Uni-Bonn

Parameter estimation and forecasting. Cristiano Porciani AIfA, Uni-Bonn Parameter estimation and forecasting Cristiano Porciani AIfA, Uni-Bonn Questions? C. Porciani Estimation & forecasting 2 Temperature fluctuations Variance at multipole l (angle ~180o/l) C. Porciani Estimation

More information

STA 294: Stochastic Processes & Bayesian Nonparametrics

STA 294: Stochastic Processes & Bayesian Nonparametrics MARKOV CHAINS AND CONVERGENCE CONCEPTS Markov chains are among the simplest stochastic processes, just one step beyond iid sequences of random variables. Traditionally they ve been used in modelling a

More information

April 20th, Advanced Topics in Machine Learning California Institute of Technology. Markov Chain Monte Carlo for Machine Learning

April 20th, Advanced Topics in Machine Learning California Institute of Technology. Markov Chain Monte Carlo for Machine Learning for for Advanced Topics in California Institute of Technology April 20th, 2017 1 / 50 Table of Contents for 1 2 3 4 2 / 50 History of methods for Enrico Fermi used to calculate incredibly accurate predictions

More information

Markov Chain Monte Carlo methods

Markov Chain Monte Carlo methods Markov Chain Monte Carlo methods By Oleg Makhnin 1 Introduction a b c M = d e f g h i 0 f(x)dx 1.1 Motivation 1.1.1 Just here Supresses numbering 1.1.2 After this 1.2 Literature 2 Method 2.1 New math As

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

Bayesian model selection in graphs by using BDgraph package

Bayesian model selection in graphs by using BDgraph package Bayesian model selection in graphs by using BDgraph package A. Mohammadi and E. Wit March 26, 2013 MOTIVATION Flow cytometry data with 11 proteins from Sachs et al. (2005) RESULT FOR CELL SIGNALING DATA

More information

Lecture 6: Markov Chain Monte Carlo

Lecture 6: Markov Chain Monte Carlo Lecture 6: Markov Chain Monte Carlo D. Jason Koskinen koskinen@nbi.ku.dk Photo by Howard Jackman University of Copenhagen Advanced Methods in Applied Statistics Feb - Apr 2016 Niels Bohr Institute 2 Outline

More information

Bayesian Image Segmentation Using MRF s Combined with Hierarchical Prior Models

Bayesian Image Segmentation Using MRF s Combined with Hierarchical Prior Models Bayesian Image Segmentation Using MRF s Combined with Hierarchical Prior Models Kohta Aoki 1 and Hiroshi Nagahashi 2 1 Interdisciplinary Graduate School of Science and Engineering, Tokyo Institute of Technology

More information

Bayesian Methods with Monte Carlo Markov Chains II

Bayesian Methods with Monte Carlo Markov Chains II Bayesian Methods with Monte Carlo Markov Chains II Henry Horng-Shing Lu Institute of Statistics National Chiao Tung University hslu@stat.nctu.edu.tw http://tigpbp.iis.sinica.edu.tw/courses.htm 1 Part 3

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

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

Multicanonical methods

Multicanonical methods Multicanonical methods Normal Monte Carlo algorithms sample configurations with the Boltzmann weight p exp( βe). Sometimes this is not desirable. Example: if a system has a first order phase transitions

More information

Winter 2019 Math 106 Topics in Applied Mathematics. Lecture 9: Markov Chain Monte Carlo

Winter 2019 Math 106 Topics in Applied Mathematics. Lecture 9: Markov Chain Monte Carlo Winter 2019 Math 106 Topics in Applied Mathematics Data-driven Uncertainty Quantification Yoonsang Lee (yoonsang.lee@dartmouth.edu) Lecture 9: Markov Chain Monte Carlo 9.1 Markov Chain A Markov Chain Monte

More information

Machine Learning Summer School

Machine Learning Summer School Machine Learning Summer School Lecture 3: Learning parameters and structure Zoubin Ghahramani zoubin@eng.cam.ac.uk http://learning.eng.cam.ac.uk/zoubin/ Department of Engineering University of Cambridge,

More information

Calibration of Stochastic Volatility Models using Particle Markov Chain Monte Carlo Methods

Calibration of Stochastic Volatility Models using Particle Markov Chain Monte Carlo Methods Calibration of Stochastic Volatility Models using Particle Markov Chain Monte Carlo Methods Jonas Hallgren 1 1 Department of Mathematics KTH Royal Institute of Technology Stockholm, Sweden BFS 2012 June

More information

Stochastic Approximation Monte Carlo and Its Applications

Stochastic Approximation Monte Carlo and Its Applications Stochastic Approximation Monte Carlo and Its Applications Faming Liang Department of Statistics Texas A&M University 1. Liang, F., Liu, C. and Carroll, R.J. (2007) Stochastic approximation in Monte Carlo

More information

an introduction to bayesian inference

an introduction to bayesian inference with an application to network analysis http://jakehofman.com january 13, 2010 motivation would like models that: provide predictive and explanatory power are complex enough to describe observed phenomena

More information

Markov Chain Monte Carlo

Markov Chain Monte Carlo Markov Chain Monte Carlo (and Bayesian Mixture Models) David M. Blei Columbia University October 14, 2014 We have discussed probabilistic modeling, and have seen how the posterior distribution is the critical

More information