Model-Based Dynamic Sampling (MBDS) in 2D and 3D. Dilshan Godaliyadda Gregery Buzzard Charles Bouman Purdue University

Size: px
Start display at page:

Download "Model-Based Dynamic Sampling (MBDS) in 2D and 3D. Dilshan Godaliyadda Gregery Buzzard Charles Bouman Purdue University"

Transcription

1 Model-Based Dynamic Sampling (MBDS) in 2D and 3D Dilshan Godaliyadda Gregery Buzzard Charles Bouman Purdue University

2 2D MBDS

3 Raster Sampling: Optimally Bad Each new sample provides the least information.

4 Random Sampling: Better Each new sample provides much more information.

5 Optimal (Greedy) Sampling: Best Each new sample provides the most information.

6 How to Select a New Measurement First model the posterior distribution p ( k x y ) k, with a Gaussian p k ( x y ) k = 1 z exp # 1 2 y 2 % k A k x $ Λk 1 & 2 xt Bx( ' From Forward Model From Prior Model New measurement location The pixel location where the posterior variance is largest

7 Recursion for Optimal Greedy Sampling For each new sample { y k = A k x + w k Step 1: Measure signal µ k = Ε x y k R k = Ε ( x µ k ) x µ k m k argmax m M ( ) T y k ( mt R k m) Step 2: Find posterior covariance of image When everything is Gaussian, R k does not depend on y k! Step 3: Select pixel with largest variance A k = A k m k Step 4: Add new row to A matrix }

8 Non-Gaussian Prior => Dynamic Sampling x 1 best measurement direc:on Likely loca:on of solu:on x 2 prior manifold

9 Non-Gaussian Prior => Dynamic Sampling best measurement direc:on x 1 Likely loca:on of solu:on x 2 prior manifold Optimal sampling depends dynamically on previous samples

10 Non-Gaussian Posterior Hence we model the posterior distribution, with a non-gaussian p k ( x y ) k p k ( x y k ) = 1 z exp 1 2 y A x 2 k k Λ k 1 2 {i, j} P x (i) x ( j ) σ c + x(i) x ( j ) σ q q p Same Forward Model Non-Gaussian Prior Model Q-GGMRF All neighboring pairs of pixels

11 2D-MBDS Algorithm For each new sample { y k = A k x + w k Step 1: Measure signal How do you generate samples from the posterior distribution? { x k (1), x k (2 ),!, x k ( L ) } ~ p( x y k ) ˆµ k (n) 1 L i=1 (n) x k (l ) n k = argmax n n (n) ( σˆ k ) 2 1 L ( )( ) T x (n) L 1 ˆµ (n) k (l ) x (n) (n) k k (l ) ˆµ k i=1 (n) ( σˆ k ) 2 Step 2: Generate L samples from posterior distribution Step 3: Estimate posterior variance Step 4: Select pixel with largest variance A k+1 = A k 0,,1,,0 Step 5: Add new row to A matrix n k }

12 Sampling from the Posterior Distribution The image below is the MAP es:mate aber the k th itera:on The posterior distribu:on Want sample images from this distribu:on Pick a Monte- Carlo Sampling method p k ( x y k ) = 1 z exp 1 2 y A x 2 k k 1 Λ k 2 {i, j} P x (i) x ( j ) σ c + x(i) x ( j ) σ q q p sample images Vectors from this distribu:on Use Has:ngs- Metropolis (HM) Algorithm Energy - u k (x) HM Algorithm Start with a good guess for a sample vector (can use the MAP es:mate itself) X 1 Draw a sample vector from a proposal distribu:on W ~ q k w X i ( ) Our proposal distribu:on A Gaussian approxima:on to p k (x y k ) Why? Because easy to draw samples from a Gaussian Calculate " $ α = min 1, q k ( X i W ) # %$ q k W X i ( ) exp { [ u(w ) u(x i) ]} & $ ' ($ For i =1:M How? By calcula:ng the second order Taylor series approxima:on to u k (x) at the MAP es:mate " W with probability α X i+1 # $ X i with probability 1-α INTRACTABLE FOR MOST IMAGES!

13 Posterior Sampling for Images (2D) Compute posterior variance for each voxel loca:on Find pixel loca:on with largest posterior variance (index n k ) { x k (1), x k (2 ),!, x k ( L ) } MAP es:ma:on to reconstruct block n k Formulate Proposal distribu:on for Has:ngs- Metropolis Sampler ˆx k Proposal Distribution for HM Sampler Generate block sample vector using Has:ngs- Metropolis Sampler Extract blocks surrounding measured voxel loca:on from the l th sample image x Generate L Block Sample Images x k (l ) Do 5ll enough measurements are taken x k (1), x k (2 ),...x k ( L )

14 Dynamic Sampling for BSE Image x 200 Segment (Image from Georgia Tech Team) 45% of image sampled Original Image Selected Samples Measured Image Reconstructed Image

15 2D MBDS With a Distance Penalty

16 Incorporating a Distance Penalty For each new sample { y k = A k x + w k { x k (1), x k (2 ),!, x k ( L ) } ~ p( x y k ) ˆµ k (n) 1 L n i=1 (n) ( σˆ k ) 2 1 (n) x k (l ) L ( )( ) T x (n) L 1 ˆµ (n) k (l ) x (n) (n) k k (l ) ˆµ k i=1 Step 1: Measure signal Step 2: Generate L samples from posterior Step 3: Estimate posterior variance n k = max n (n) ( σ k ) 2 Δt (i,s) where Δt (i,s) = α + βd (i,s) Step 4: Select pixel with largest variance according to the distance penalty d (i,s) α β - distance between location s and i. - time taken to acquire a measurement after beam is moved - time taken to move beam a unit distance A k+1 = A k 0,,1,,0 n (k) Step 5: Add new row to A matrix }

17 Incorporating a Distance Penalty Reconstructed Images 5% 10% 15% 20% d 0 =50 (Penalized rela:vely more for distance) d 0 =10,000 (Penalized rela:vely less for distance)

18 3D MBDS

19 3D - MBDS Algorithm For each slice (for t = 1:T){ For each new measurement (for k=1:k){ y k,t = A k,t x t + w k,t { ( )} ˆx t argmax p x t y t,k, y t,k, x t i for i = 1,2,3... x t Step 1: Measure Image Step 2: Estimate the image using MAP Estimation { x, x,!, x t (1),k t (2 ),k t ( L ),k} ~ p( x t y t,k, y t i,k, x t i for i = 1,2,3... ) ˆµ (n) t,k 1 L L i=1 x (n) t (i ),k ( σ ˆ (n) t,k ) 2 1 L 1 L ( x (n) ˆµ (n) t (i ),k t,k ) x (n) i=1 ( ˆµ (n) t (i ),k t,k ) T Step 3: Generate L samples from posterior y t i,k x t i - Measurements from previous slices - Previous slices Step 4: Estimate posterior variance n t,k = max n (n) ( σ t,k ) 2 Δt (i,s) where Δt (i,s) = α + βd (i,s) Step 5: Select voxel with largest variance } A t,k+1 = } A t,k 0,,1,,0 n t,k Step 6: Add new row to A t matrix

20 Posterior Sampling for Images (3D) Compute posterior variance for each voxel loca:on Find pixel loca:on with largest posterior variance (index n t,k ) ˆx t 1,K, ˆx t 2,K, ˆx t 3,K { x t,k 1 (1), x t,k 1 (2 ),!, x t,k 1 ( L ) } n t,k Formulate Proposal distribu:on for Has:ngs- Metropolis Sampler Proposal Distribution for HM Sampler Generate block sample vector using Has:ngs- Metropolis Sampler x t,k (l ) Extract blocks surrounding measured voxel loca:on from the l x th sample image x t 1,K (l ), x t 2,K (l ), x t 3,K (l ) Do 5ll enough measurements are taken MAP es:ma:on to reconstruct block ˆx t,k Generate L Block Sample Images x t,k (1), x t,k (2 ),...x t,k ( L )

21 The Setup for the experiment Temporal neighborhood (z-direction) = 1 Spacial neighborhood (x-y plane) = 1 z x Need a condition to stop measuring a slice and move on to the next one y We use a threshold on Expected Variance For the experiment the threshold used is 0.3 This means for any voxel in the 1 st time slice we consider 8 neighbors and for all the others we consider 17 neighbors.

22 Results Reconstructions when Sampling Terminates Slice 0 (x 0 ) Slice 1 (x 1 ) Slice 2 (x 2 ) Slice 3 (x 3 ) Reconstructed Image Sampled locations Measured Image Original Images Percentage when sampling terminated for each slice ~64% ~34% ~40% ~33%

Physics-based Prior modeling in Inverse Problems

Physics-based Prior modeling in Inverse Problems Physics-based Prior modeling in Inverse Problems MURI Meeting 2013 M Usman Sadiq, Purdue University Charles A. Bouman, Purdue University In collaboration with: Jeff Simmons, AFRL Venkat Venkatakrishnan,

More information

F denotes cumulative density. denotes probability density function; (.)

F denotes cumulative density. denotes probability density function; (.) BAYESIAN ANALYSIS: FOREWORDS Notation. System means the real thing and a model is an assumed mathematical form for the system.. he probability model class M contains the set of the all admissible models

More information

CSE446: Linear Regression Regulariza5on Bias / Variance Tradeoff Winter 2015

CSE446: Linear Regression Regulariza5on Bias / Variance Tradeoff Winter 2015 CSE446: Linear Regression Regulariza5on Bias / Variance Tradeoff Winter 2015 Luke ZeElemoyer Slides adapted from Carlos Guestrin Predic5on of con5nuous variables Billionaire says: Wait, that s not what

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

Par$cle Filters Part I: Theory. Peter Jan van Leeuwen Data- Assimila$on Research Centre DARC University of Reading

Par$cle Filters Part I: Theory. Peter Jan van Leeuwen Data- Assimila$on Research Centre DARC University of Reading Par$cle Filters Part I: Theory Peter Jan van Leeuwen Data- Assimila$on Research Centre DARC University of Reading Reading July 2013 Why Data Assimila$on Predic$on Model improvement: - Parameter es$ma$on

More information

Nonlinear ensemble data assimila/on in high- dimensional spaces. Peter Jan van Leeuwen Javier Amezcua, Mengbin Zhu, Melanie Ades

Nonlinear ensemble data assimila/on in high- dimensional spaces. Peter Jan van Leeuwen Javier Amezcua, Mengbin Zhu, Melanie Ades Nonlinear ensemble data assimila/on in high- dimensional spaces Peter Jan van Leeuwen Javier Amezcua, Mengbin Zhu, Melanie Ades Data assimila/on: general formula/on Bayes theorem: The solu/on is a pdf!

More information

Continuous State MRF s

Continuous State MRF s EE64 Digital Image Processing II: Purdue University VISE - December 4, Continuous State MRF s Topics to be covered: Quadratic functions Non-Convex functions Continuous MAP estimation Convex functions EE64

More information

CS 6140: Machine Learning Spring What We Learned Last Week. Survey 2/26/16. VS. Model

CS 6140: Machine Learning Spring What We Learned Last Week. Survey 2/26/16. VS. Model Logis@cs CS 6140: Machine Learning Spring 2016 Instructor: Lu Wang College of Computer and Informa@on Science Northeastern University Webpage: www.ccs.neu.edu/home/luwang Email: luwang@ccs.neu.edu Assignment

More information

CS 6140: Machine Learning Spring 2016

CS 6140: Machine Learning Spring 2016 CS 6140: Machine Learning Spring 2016 Instructor: Lu Wang College of Computer and Informa?on Science Northeastern University Webpage: www.ccs.neu.edu/home/luwang Email: luwang@ccs.neu.edu Logis?cs Assignment

More information

Introduc)on to Ar)ficial Intelligence

Introduc)on to Ar)ficial Intelligence Introduc)on to Ar)ficial Intelligence Lecture 13 Approximate Inference CS/CNS/EE 154 Andreas Krause Bayesian networks! Compact representa)on of distribu)ons over large number of variables! (OQen) allows

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

An Introduction to Expectation-Maximization

An Introduction to Expectation-Maximization An Introduction to Expectation-Maximization Dahua Lin Abstract This notes reviews the basics about the Expectation-Maximization EM) algorithm, a popular approach to perform model estimation of the generative

More information

Bayesian Inference for DSGE Models. Lawrence J. Christiano

Bayesian Inference for DSGE Models. Lawrence J. Christiano Bayesian Inference for DSGE Models Lawrence J. Christiano Outline State space-observer form. convenient for model estimation and many other things. Preliminaries. Probabilities. Maximum Likelihood. Bayesian

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

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

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

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

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

Drawing Inferences from Statistics Based on Multiyear Asset Returns

Drawing Inferences from Statistics Based on Multiyear Asset Returns Drawing Inferences from Statistics Based on Multiyear Asset Returns Matthew Richardson ames H. Stock FE 1989 1 Motivation Fama and French (1988, Poterba and Summer (1988 document significant negative correlations

More information

Bayesian linear regression

Bayesian linear regression Bayesian linear regression Linear regression is the basis of most statistical modeling. The model is Y i = X T i β + ε i, where Y i is the continuous response X i = (X i1,..., X ip ) T is the corresponding

More information

CSC 2541: Bayesian Methods for Machine Learning

CSC 2541: Bayesian Methods for Machine Learning CSC 2541: Bayesian Methods for Machine Learning Radford M. Neal, University of Toronto, 2011 Lecture 3 More Markov Chain Monte Carlo Methods The Metropolis algorithm isn t the only way to do MCMC. We ll

More information

Unsupervised Learning: K- Means & PCA

Unsupervised Learning: K- Means & PCA Unsupervised Learning: K- Means & PCA Unsupervised Learning Supervised learning used labeled data pairs (x, y) to learn a func>on f : X Y But, what if we don t have labels? No labels = unsupervised learning

More information

Online appendix to On the stability of the excess sensitivity of aggregate consumption growth in the US

Online appendix to On the stability of the excess sensitivity of aggregate consumption growth in the US Online appendix to On the stability of the excess sensitivity of aggregate consumption growth in the US Gerdie Everaert 1, Lorenzo Pozzi 2, and Ruben Schoonackers 3 1 Ghent University & SHERPPA 2 Erasmus

More information

Linear inverse Gaussian theory and geostatistics a tomography example København Ø,

Linear inverse Gaussian theory and geostatistics a tomography example København Ø, Linear inverse Gaussian theory and geostatistics a tomography example Thomas Mejer Hansen 1, ndre Journel 2, lbert Tarantola 3 and Klaus Mosegaard 1 1 Niels Bohr Institute, University of Copenhagen, Juliane

More information

Regulatory Inferece from Gene Expression. CMSC858P Spring 2012 Hector Corrada Bravo

Regulatory Inferece from Gene Expression. CMSC858P Spring 2012 Hector Corrada Bravo Regulatory Inferece from Gene Expression CMSC858P Spring 2012 Hector Corrada Bravo 2 Graphical Model Let y be a vector- valued random variable Suppose some condi8onal independence proper8es hold for some

More information

Bayesian Inference for DSGE Models. Lawrence J. Christiano

Bayesian Inference for DSGE Models. Lawrence J. Christiano Bayesian Inference for DSGE Models Lawrence J. Christiano Outline State space-observer form. convenient for model estimation and many other things. Bayesian inference Bayes rule. Monte Carlo integation.

More information

Machine Learning and Data Mining. Bayes Classifiers. Prof. Alexander Ihler

Machine Learning and Data Mining. Bayes Classifiers. Prof. Alexander Ihler + Machine Learning and Data Mining Bayes Classifiers Prof. Alexander Ihler A basic classifier Training data D={x (i),y (i) }, Classifier f(x ; D) Discrete feature vector x f(x ; D) is a con@ngency table

More information

Algebra of Random Variables: Optimal Average and Optimal Scaling Minimising

Algebra of Random Variables: Optimal Average and Optimal Scaling Minimising Review: Optimal Average/Scaling is equivalent to Minimise χ Two 1-parameter models: Estimating < > : Scaling a pattern: Two equivalent methods: Algebra of Random Variables: Optimal Average and Optimal

More information

Algebra of Random Variables: Optimal Average and Optimal Scaling Minimising

Algebra of Random Variables: Optimal Average and Optimal Scaling Minimising Review: Optimal Average/Scaling is equivalent to Minimise χ Two 1-parameter models: Estimating < > : Scaling a pattern: Two equivalent methods: Algebra of Random Variables: Optimal Average and Optimal

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

Sequential Monte Carlo and Particle Filtering. Frank Wood Gatsby, November 2007

Sequential Monte Carlo and Particle Filtering. Frank Wood Gatsby, November 2007 Sequential Monte Carlo and Particle Filtering Frank Wood Gatsby, November 2007 Importance Sampling Recall: Let s say that we want to compute some expectation (integral) E p [f] = p(x)f(x)dx and we remember

More information

DAG models and Markov Chain Monte Carlo methods a short overview

DAG models and Markov Chain Monte Carlo methods a short overview DAG models and Markov Chain Monte Carlo methods a short overview Søren Højsgaard Institute of Genetics and Biotechnology University of Aarhus August 18, 2008 Printed: August 18, 2008 File: DAGMC-Lecture.tex

More information

Introduction to Particle Filters for Data Assimilation

Introduction to Particle Filters for Data Assimilation Introduction to Particle Filters for Data Assimilation Mike Dowd Dept of Mathematics & Statistics (and Dept of Oceanography Dalhousie University, Halifax, Canada STATMOS Summer School in Data Assimila5on,

More information

DART Tutorial Sec'on 1: Filtering For a One Variable System

DART Tutorial Sec'on 1: Filtering For a One Variable System DART Tutorial Sec'on 1: Filtering For a One Variable System UCAR The Na'onal Center for Atmospheric Research is sponsored by the Na'onal Science Founda'on. Any opinions, findings and conclusions or recommenda'ons

More information

Clustering K-means. Machine Learning CSE546. Sham Kakade University of Washington. November 15, Review: PCA Start: unsupervised learning

Clustering K-means. Machine Learning CSE546. Sham Kakade University of Washington. November 15, Review: PCA Start: unsupervised learning Clustering K-means Machine Learning CSE546 Sham Kakade University of Washington November 15, 2016 1 Announcements: Project Milestones due date passed. HW3 due on Monday It ll be collaborative HW2 grades

More information

CS 4495 Computer Vision Principle Component Analysis

CS 4495 Computer Vision Principle Component Analysis CS 4495 Computer Vision Principle Component Analysis (and it s use in Computer Vision) Aaron Bobick School of Interactive Computing Administrivia PS6 is out. Due *** Sunday, Nov 24th at 11:55pm *** PS7

More information

Markov chain Monte Carlo methods in atmospheric remote sensing

Markov chain Monte Carlo methods in atmospheric remote sensing 1 / 45 Markov chain Monte Carlo methods in atmospheric remote sensing Johanna Tamminen johanna.tamminen@fmi.fi ESA Summer School on Earth System Monitoring and Modeling July 3 Aug 11, 212, Frascati July,

More information

Clustering K-means. Clustering images. Machine Learning CSE546 Carlos Guestrin University of Washington. November 4, 2014.

Clustering K-means. Clustering images. Machine Learning CSE546 Carlos Guestrin University of Washington. November 4, 2014. Clustering K-means Machine Learning CSE546 Carlos Guestrin University of Washington November 4, 2014 1 Clustering images Set of Images [Goldberger et al.] 2 1 K-means Randomly initialize k centers µ (0)

More information

Covariance Matrix Simplification For Efficient Uncertainty Management

Covariance Matrix Simplification For Efficient Uncertainty Management PASEO MaxEnt 2007 Covariance Matrix Simplification For Efficient Uncertainty Management André Jalobeanu, Jorge A. Gutiérrez PASEO Research Group LSIIT (CNRS/ Univ. Strasbourg) - Illkirch, France *part

More information

Advances and Applications in Perfect Sampling

Advances and Applications in Perfect Sampling and Applications in Perfect Sampling Ph.D. Dissertation Defense Ulrike Schneider advisor: Jem Corcoran May 8, 2003 Department of Applied Mathematics University of Colorado Outline Introduction (1) MCMC

More information

Spatial Lasso with Application to GIS Model Selection. F. Jay Breidt Colorado State University

Spatial Lasso with Application to GIS Model Selection. F. Jay Breidt Colorado State University Spatial Lasso with Application to GIS Model Selection F. Jay Breidt Colorado State University with Hsin-Cheng Huang, Nan-Jung Hsu, and Dave Theobald September 25 The work reported here was developed under

More information

Bayesian Methods for Sparse Signal Recovery

Bayesian Methods for Sparse Signal Recovery Bayesian Methods for Sparse Signal Recovery Bhaskar D Rao 1 University of California, San Diego 1 Thanks to David Wipf, Jason Palmer, Zhilin Zhang and Ritwik Giri Motivation Motivation Sparse Signal Recovery

More information

Forward Problems and their Inverse Solutions

Forward Problems and their Inverse Solutions Forward Problems and their Inverse Solutions Sarah Zedler 1,2 1 King Abdullah University of Science and Technology 2 University of Texas at Austin February, 2013 Outline 1 Forward Problem Example Weather

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

1. Introduc9on 2. Bivariate Data 3. Linear Analysis of Data

1. Introduc9on 2. Bivariate Data 3. Linear Analysis of Data Lecture 3: Bivariate Data & Linear Regression 1. Introduc9on 2. Bivariate Data 3. Linear Analysis of Data a) Freehand Linear Fit b) Least Squares Fit c) Interpola9on/Extrapola9on 4. Correla9on 1. Introduc9on

More information

CSE 473: Ar+ficial Intelligence. Example. Par+cle Filters for HMMs. An HMM is defined by: Ini+al distribu+on: Transi+ons: Emissions:

CSE 473: Ar+ficial Intelligence. Example. Par+cle Filters for HMMs. An HMM is defined by: Ini+al distribu+on: Transi+ons: Emissions: CSE 473: Ar+ficial Intelligence Par+cle Filters for HMMs Daniel S. Weld - - - University of Washington [Most slides were created by Dan Klein and Pieter Abbeel for CS188 Intro to AI at UC Berkeley. All

More information

Bellman s Curse of Dimensionality

Bellman s Curse of Dimensionality Bellman s Curse of Dimensionality n- dimensional state space Number of states grows exponen

More information

Physics 403. Segev BenZvi. Numerical Methods, Maximum Likelihood, and Least Squares. Department of Physics and Astronomy University of Rochester

Physics 403. Segev BenZvi. Numerical Methods, Maximum Likelihood, and Least Squares. Department of Physics and Astronomy University of Rochester Physics 403 Numerical Methods, Maximum Likelihood, and Least Squares Segev BenZvi Department of Physics and Astronomy University of Rochester Table of Contents 1 Review of Last Class Quadratic Approximation

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

Nonparametric Drift Estimation for Stochastic Differential Equations

Nonparametric Drift Estimation for Stochastic Differential Equations Nonparametric Drift Estimation for Stochastic Differential Equations Gareth Roberts 1 Department of Statistics University of Warwick Brazilian Bayesian meeting, March 2010 Joint work with O. Papaspiliopoulos,

More information

CS4670: Computer Vision Kavita Bala. Lecture 7: Harris Corner Detec=on

CS4670: Computer Vision Kavita Bala. Lecture 7: Harris Corner Detec=on CS4670: Computer Vision Kavita Bala Lecture 7: Harris Corner Detec=on Announcements HW 1 will be out soon Sign up for demo slots for PA 1 Remember that both partners have to be there We will ask you to

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

Spatial Statistics with Image Analysis. Outline. A Statistical Approach. Johan Lindström 1. Lund October 6, 2016

Spatial Statistics with Image Analysis. Outline. A Statistical Approach. Johan Lindström 1. Lund October 6, 2016 Spatial Statistics Spatial Examples More Spatial Statistics with Image Analysis Johan Lindström 1 1 Mathematical Statistics Centre for Mathematical Sciences Lund University Lund October 6, 2016 Johan Lindström

More information

Strong Lens Modeling (II): Statistical Methods

Strong Lens Modeling (II): Statistical Methods Strong Lens Modeling (II): Statistical Methods Chuck Keeton Rutgers, the State University of New Jersey Probability theory multiple random variables, a and b joint distribution p(a, b) conditional distribution

More information

Markov chain Monte Carlo Lecture 9

Markov chain Monte Carlo Lecture 9 Markov chain Monte Carlo Lecture 9 David Sontag New York University Slides adapted from Eric Xing and Qirong Ho (CMU) Limitations of Monte Carlo Direct (unconditional) sampling Hard to get rare events

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

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

Bayesian Gaussian / Linear Models. Read Sections and 3.3 in the text by Bishop

Bayesian Gaussian / Linear Models. Read Sections and 3.3 in the text by Bishop Bayesian Gaussian / Linear Models Read Sections 2.3.3 and 3.3 in the text by Bishop Multivariate Gaussian Model with Multivariate Gaussian Prior Suppose we model the observed vector b as having a multivariate

More information

STA 414/2104, Spring 2014, Practice Problem Set #1

STA 414/2104, Spring 2014, Practice Problem Set #1 STA 44/4, Spring 4, Practice Problem Set # Note: these problems are not for credit, and not to be handed in Question : Consider a classification problem in which there are two real-valued inputs, and,

More information

Introduction to Machine Learning. Maximum Likelihood and Bayesian Inference. Lecturers: Eran Halperin, Yishay Mansour, Lior Wolf

Introduction to Machine Learning. Maximum Likelihood and Bayesian Inference. Lecturers: Eran Halperin, Yishay Mansour, Lior Wolf 1 Introduction to Machine Learning Maximum Likelihood and Bayesian Inference Lecturers: Eran Halperin, Yishay Mansour, Lior Wolf 2013-14 We know that X ~ B(n,p), but we do not know p. We get a random sample

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

Universität Potsdam Institut für Informatik Lehrstuhl Maschinelles Lernen PCA. Tobias Scheffer

Universität Potsdam Institut für Informatik Lehrstuhl Maschinelles Lernen PCA. Tobias Scheffer Universität Potsdam Institut für Informatik Lehrstuhl Maschinelles Lernen PCA Tobias Scheffer Overview Principal Component Analysis (PCA) Kernel-PCA Fisher Linear Discriminant Analysis t-sne 2 PCA: Motivation

More information

MID-TERM EXAM ANSWERS. p t + δ t = Rp t 1 + η t (1.1)

MID-TERM EXAM ANSWERS. p t + δ t = Rp t 1 + η t (1.1) ECO 513 Fall 2005 C.Sims MID-TERM EXAM ANSWERS (1) Suppose a stock price p t and the stock dividend δ t satisfy these equations: p t + δ t = Rp t 1 + η t (1.1) δ t = γδ t 1 + φp t 1 + ε t, (1.2) where

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

Slides modified from: PATTERN RECOGNITION AND MACHINE LEARNING CHRISTOPHER M. BISHOP

Slides modified from: PATTERN RECOGNITION AND MACHINE LEARNING CHRISTOPHER M. BISHOP Slides modified from: PATTERN RECOGNITION AND MACHINE LEARNING CHRISTOPHER M. BISHOP Predic?ve Distribu?on (1) Predict t for new values of x by integra?ng over w: where The Evidence Approxima?on (1) The

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

Bayesian Inference for DSGE Models. Lawrence J. Christiano

Bayesian Inference for DSGE Models. Lawrence J. Christiano Bayesian Inference for DSGE Models Lawrence J. Christiano Outline State space-observer form. convenient for model estimation and many other things. Bayesian inference Bayes rule. Monte Carlo integation.

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

Harmonic oscillator. U(x) = 1 2 bx2

Harmonic oscillator. U(x) = 1 2 bx2 Harmonic oscillator The harmonic oscillator is a familiar problem from classical mechanics. The situation is described by a force which depends linearly on distance as happens with the restoring force

More information

Linear Regression and Correla/on. Correla/on and Regression Analysis. Three Ques/ons 9/14/14. Chapter 13. Dr. Richard Jerz

Linear Regression and Correla/on. Correla/on and Regression Analysis. Three Ques/ons 9/14/14. Chapter 13. Dr. Richard Jerz Linear Regression and Correla/on Chapter 13 Dr. Richard Jerz 1 Correla/on and Regression Analysis Correla/on Analysis is the study of the rela/onship between variables. It is also defined as group of techniques

More information

Linear Regression and Correla/on

Linear Regression and Correla/on Linear Regression and Correla/on Chapter 13 Dr. Richard Jerz 1 Correla/on and Regression Analysis Correla/on Analysis is the study of the rela/onship between variables. It is also defined as group of techniques

More information

Penalty and Barrier Methods General classical constrained minimization problem minimize f(x) subject to g(x) 0 h(x) =0 Penalty methods are motivated by the desire to use unconstrained optimization techniques

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

MAP Reconstruction From Spatially Correlated PET Data

MAP Reconstruction From Spatially Correlated PET Data IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL 50, NO 5, OCTOBER 2003 1445 MAP Reconstruction From Spatially Correlated PET Data Adam Alessio, Student Member, IEEE, Ken Sauer, Member, IEEE, and Charles A Bouman,

More information

Results: MCMC Dancers, q=10, n=500

Results: MCMC Dancers, q=10, n=500 Motivation Sampling Methods for Bayesian Inference How to track many INTERACTING targets? A Tutorial Frank Dellaert Results: MCMC Dancers, q=10, n=500 1 Probabilistic Topological Maps Results Real-Time

More information

16 : Approximate Inference: Markov Chain Monte Carlo

16 : Approximate Inference: Markov Chain Monte Carlo 10-708: Probabilistic Graphical Models 10-708, Spring 2017 16 : Approximate Inference: Markov Chain Monte Carlo Lecturer: Eric P. Xing Scribes: Yuan Yang, Chao-Ming Yen 1 Introduction As the target distribution

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

arxiv: v1 [stat.co] 18 Feb 2012

arxiv: v1 [stat.co] 18 Feb 2012 A LEVEL-SET HIT-AND-RUN SAMPLER FOR QUASI-CONCAVE DISTRIBUTIONS Dean Foster and Shane T. Jensen arxiv:1202.4094v1 [stat.co] 18 Feb 2012 Department of Statistics The Wharton School University of Pennsylvania

More information

Econ 2148, fall 2017 Gaussian process priors, reproducing kernel Hilbert spaces, and Splines

Econ 2148, fall 2017 Gaussian process priors, reproducing kernel Hilbert spaces, and Splines Econ 2148, fall 2017 Gaussian process priors, reproducing kernel Hilbert spaces, and Splines Maximilian Kasy Department of Economics, Harvard University 1 / 37 Agenda 6 equivalent representations of the

More information

Statistical Pattern Recognition

Statistical Pattern Recognition Statistical Pattern Recognition Expectation Maximization (EM) and Mixture Models Hamid R. Rabiee Jafar Muhammadi, Mohammad J. Hosseini Spring 2014 http://ce.sharif.edu/courses/92-93/2/ce725-2 Agenda Expectation-maximization

More information

Machine Learning, Fall 2009: Midterm

Machine Learning, Fall 2009: Midterm 10-601 Machine Learning, Fall 009: Midterm Monday, November nd hours 1. Personal info: Name: Andrew account: E-mail address:. You are permitted two pages of notes and a calculator. Please turn off all

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

Bayesian data analysis in practice: Three simple examples

Bayesian data analysis in practice: Three simple examples Bayesian data analysis in practice: Three simple examples Martin P. Tingley Introduction These notes cover three examples I presented at Climatea on 5 October 0. Matlab code is available by request to

More information

Tangent lines, cont d. Linear approxima5on and Newton s Method

Tangent lines, cont d. Linear approxima5on and Newton s Method Tangent lines, cont d Linear approxima5on and Newton s Method Last 5me: A challenging tangent line problem, because we had to figure out the point of tangency.?? (A) I get it! (B) I think I see how we

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

Lecture 6: Multiple Model Filtering, Particle Filtering and Other Approximations

Lecture 6: Multiple Model Filtering, Particle Filtering and Other Approximations Lecture 6: Multiple Model Filtering, Particle Filtering and Other Approximations Department of Biomedical Engineering and Computational Science Aalto University April 28, 2010 Contents 1 Multiple Model

More information

Pattern Recognition and Machine Learning

Pattern Recognition and Machine Learning Christopher M. Bishop Pattern Recognition and Machine Learning ÖSpri inger Contents Preface Mathematical notation Contents vii xi xiii 1 Introduction 1 1.1 Example: Polynomial Curve Fitting 4 1.2 Probability

More information

Numerical Methods in Physics

Numerical Methods in Physics Numerical Methods in Physics Numerische Methoden in der Physik, 515.421. Instructor: Ass. Prof. Dr. Lilia Boeri Room: PH 03 090 Tel: +43-316- 873 8191 Email Address: l.boeri@tugraz.at Room: TDK Seminarraum

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

Bayesian Complementary Clustering, MCMC and Anglo-Saxon placenames

Bayesian Complementary Clustering, MCMC and Anglo-Saxon placenames Bayesian Complementary Clustering, MCMC and Anglo-Saxon placenames Giacomo Zanella g.zanella@warwick.ac.uk Department of Statistics University of Warwick, Coventry, UK 20 May 2014 Overview 1. Motivation:

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

Implicit sampling for particle filters. Alexandre Chorin, Mathias Morzfeld, Xuemin Tu, Ethan Atkins

Implicit sampling for particle filters. Alexandre Chorin, Mathias Morzfeld, Xuemin Tu, Ethan Atkins 0/20 Implicit sampling for particle filters Alexandre Chorin, Mathias Morzfeld, Xuemin Tu, Ethan Atkins University of California at Berkeley 2/20 Example: Try to find people in a boat in the middle of

More information

Bayesian Machine Learning

Bayesian Machine Learning Bayesian Machine Learning Andrew Gordon Wilson ORIE 6741 Lecture 2: Bayesian Basics https://people.orie.cornell.edu/andrew/orie6741 Cornell University August 25, 2016 1 / 17 Canonical Machine Learning

More information

DART Tutorial Sec'on 5: Comprehensive Filtering Theory: Non-Iden'ty Observa'ons and the Joint Phase Space

DART Tutorial Sec'on 5: Comprehensive Filtering Theory: Non-Iden'ty Observa'ons and the Joint Phase Space DART Tutorial Sec'on 5: Comprehensive Filtering Theory: Non-Iden'ty Observa'ons and the Joint Phase Space UCAR The Na'onal Center for Atmospheric Research is sponsored by the Na'onal Science Founda'on.

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

Today: Fundamentals of Monte Carlo

Today: Fundamentals of Monte Carlo Today: Fundamentals of Monte Carlo What is Monte Carlo? Named at Los Alamos in 1940 s after the casino. Any method which uses (pseudo)random numbers as an essential part of the algorithm. Stochastic -

More information

Advanced Statistical Computing

Advanced Statistical Computing Advanced Statistical Computing Fall 206 Steve Qin Outline Collapsing, predictive updating Sequential Monte Carlo 2 Collapsing and grouping Want to sample from = Regular Gibbs sampler: Sample t+ from π

More information

Factor Analysis and Kalman Filtering (11/2/04)

Factor Analysis and Kalman Filtering (11/2/04) CS281A/Stat241A: Statistical Learning Theory Factor Analysis and Kalman Filtering (11/2/04) Lecturer: Michael I. Jordan Scribes: Byung-Gon Chun and Sunghoon Kim 1 Factor Analysis Factor analysis is used

More information

AMS 147 Computational Methods and Applications Lecture 17 Copyright by Hongyun Wang, UCSC

AMS 147 Computational Methods and Applications Lecture 17 Copyright by Hongyun Wang, UCSC Lecture 17 Copyright by Hongyun Wang, UCSC Recap: Solving linear system A x = b Suppose we are given the decomposition, A = L U. We solve (LU) x = b in 2 steps: *) Solve L y = b using the forward substitution

More information

Marginal density. If the unknown is of the form x = (x 1, x 2 ) in which the target of investigation is x 1, a marginal posterior density

Marginal density. If the unknown is of the form x = (x 1, x 2 ) in which the target of investigation is x 1, a marginal posterior density Marginal density If the unknown is of the form x = x 1, x 2 ) in which the target of investigation is x 1, a marginal posterior density πx 1 y) = πx 1, x 2 y)dx 2 = πx 2 )πx 1 y, x 2 )dx 2 needs to be

More information