DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING. Inversion basics. Erkki Kyrölä Finnish Meteorological Institute
|
|
- Dominic Walsh
- 5 years ago
- Views:
Transcription
1 Inversion basics y = Kx + ε x ˆ = (K T K) 1 K T y Erkki Kyrölä Finnish Meteorological Institute Day 3 Lecture 1 Retrieval techniques - Erkki Kyrölä 1
2 Contents 1. Introduction: Measurements, models, inversion 2. Classical inversion: LSQ 3. Bayes theory 4. Monte Carlo Markov Chain 5. References Day 3 Lecture 1 Retrieval techniques - Erkki Kyrölä 2
3 Introduction Forward model target instrument data Inverse model inverse model = generalized inverse of the approximate forward model Day 3 Lecture 1 Retrieval techniques - Erkki Kyrölä 3
4 Measurements Measurements either count or scale Remote and in situ measurements Direct and indirect measurements Classical or quantum measurements Day 3 Lecture 1 Retrieval techniques - Erkki Kyrölä 4
5 Atmospheric remote measurements Atmosphere is continuously changing in time and space. No repeated measurements of the same quantity. Radiation field measurements are direct, all other measurements are indirect Measurements probe large atmospheric volume. Large averaging. Validation by in situ measurements is difficult. Day 3 Lecture 1 Retrieval techniques - Erkki Kyrölä 5
6 Forward models The true nature G(x,z) + ε z=all other pertinent variables G known (x,z known = z fix ) + ε The best forward model available. Uninteresting variables fixed. G app (x,z known = z fix ) + ε Model used in simulation G inv (x,z known = z fix ) + ε Model used for inversion Day 3 Lecture 1 Retrieval techniques - Erkki Kyrölä 6
7 Inverse problem y = K(x) + ε x = Parameters to be determined from measurements K = Forward model y = Measurements (data) ε = Noise Find best x when y is measured. Define best first. Day 3 Lecture 1 Retrieval techniques - Erkki Kyrölä 7
8 Least squares solution (LSQ) Minimize S = (y p K p (x)) 2 Distance between data and the model prediction If we have a linear problem y = Kx we get simply ˆ x = (K T K) 1 K T y Day 3 Lecture 1 Retrieval techniques - Erkki Kyrölä 8
9 A systematic basis for inversion theory is given by the Bayesian approach Model parameters are random variables Probability distribution of model parameters is retrieved A priori information is needed. This has led to many controversies about the Bayesian approach. Day 3 Lecture 1 Retrieval techniques - Erkki Kyrölä 9
10 Bayesian method P(x y)p(y) = P(y x)p(x) P(x y) = P(y x)p(x) P(y) = P(y x)p(x) P(y x)p(x)dx P(x y) = Conditional probability distribution for model parameters x given data y P(x) = A priori probability for model parameters P(y x) = Conditional pdf for data y when x given. Also called as likelihood. P(y) = The normalization. It can usually be ignored. Day 3 Lecture 1 Retrieval techniques - Erkki Kyrölä 10
11 Various point estimators for x Mean Minimum variance Maximum probability Mean = Minimum variance estimator Mean = Maximum probability if pdf symmetric Day 3 Lecture 1 Retrieval techniques - Erkki Kyrölä 11
12 P(x y) = P(y x)p(x) Whole distribution Point estimation MCMC method max of max of Maximum likelihood MAP Gaussian errors LSQ LM method Linear model Closed solution Day 3 Lecture 1 Retrieval techniques - Erkki Kyrölä 12
13 Example: Linear problem y = K true x true + ε Kx + ε Assume Gaussian noise and prior distribution: P(y x) = const e 1 2 (y Kx)T C 1 D (y Kx) const = ((2π) N det(c D )) 1 2 P(x) = const e 1 2 (x x a ) T C 1 a (x x a ) Day 3 Lecture 1 Retrieval techniques - Erkki Kyrölä 13
14 Maximum a posteriori ˆ x = (K T C D 1 K + C a 1 ) 1 (K T C D 1 y + C a 1 x a ) Interpretation: weighted mean between data and a priori Posterior distribution Model covariance P(x y) = const e 1 2 (x ˆ x )T C x 1 (x ˆ x ) C x = (K T C D 1 K + C a 1 ) 1 The solution can be written also as ˆ x = x a + C a K T (KC a K T + C D 1 ) 1 (y Kx a ) This can be viewed as an update to a priori Assimilation Day 3 Lecture 1 Retrieval techniques - Erkki Kyrölä 14
15 Properties of linear solution ˆ x = (K T C D 1 K + C a 1 ) 1 (K T C D 1 y + C a 1 x a ) = C x K T C D 1 K true x true + C x K T C D 1 ε + C x C a 1 x a Averaging kernel A = C x K T C D 1 K true ˆ x x = (A I)(x x a ) + C x K T C D 1 ε smoothing error random error Day 3 Lecture 1 Retrieval techniques - Erkki Kyrölä 15
16 Quality of retrieval Bias ˆ x = C x K T C D 1 K true x true + C x K T C D 1 ε + C x C a 1 x a linear If C a and K = K true no bias The bias in retrieval can usually be checked only by computer simulations. Intercomparisons of real measurements can also be used to detect bias. Residual Investigate the difference d = y obs y mod = y obs Kˆ x Chi2 χ 2 = 1 N m ( (y Kˆ T obs x ) 1 CD (y obs Kˆ x ) + ( x ˆ x a ) T C 1 a ( x ˆ x a )) Day 3 Lecture 1 Retrieval techniques - Erkki Kyrölä 16
17 Special cases 1. No apriori C a and K = K true A=1 and we obtain WLSQ 2. Connection to elementary data analysis. Take case 1 and only one parameter. Then the MAP estimator is mean and C x = σ 2 N i.e. the standard error of the mean. Day 3 Lecture 1 Retrieval techniques - Erkki Kyrölä 17
18 Possible solutions for linear equations Number of data = N, number of unknowns = M 1. N = M Exact inversion possible 2. N > M Overdetermined problem. Additional information may be used to constrain the solution. 3. N < M Underdetermined problem. Needs additional information or constraints Day 3 Lecture 1 Retrieval techniques - Erkki Kyrölä 18
19 Non-linear problems Gaussian errors and linear model give a quadratic optimisation problem. This leads to linear (in data) estimates. All other cases lead to non-linear problems. Gaussian errors and non-linear models can be approached by the Levenberg-Marquardt algorithm Sometimes a model can also be linearised Sometimes we can transform the problem to a new linear problem. Note: Error statistics will also change With very noisy data and/or complicated models several maxima of pdf can exist. Global methods, like simulated annealing, may help but it is better to try MCMC. Day 3 Lecture 1 Retrieval techniques - Erkki Kyrölä 19
20 Ultimate estimators: Markov chain Monte Carlo Twin peaks drama Mr. Markov: Hold your horses Blind Mr. Levenberg: That s it! Top guy: Yes! Mean guy: <Sorry but...> Flatness dullness Day 3 Lecture 1 Retrieval techniques - Erkki Kyrölä 20
21 Markov chain Monte Carlo Estimators from MCMC 1 N <x i >= Σ z t i N t Day 3 Lecture 1 Retrieval techniques - Erkki Kyrölä 21
22 MCMC examples (GOMOS) Bright star Weak star Marginal posterior distributions at 30 km for different gases Day 3 Lecture 1 Retrieval techniques - Erkki Kyrölä 22
23 A priori information Discrete grid: Assume that profile has only a finite number of free parameters Smoothness: Tikhonov constraint A priori profile Positivity constraint or similar Day 3 Lecture 1 Retrieval techniques - Erkki Kyrölä 23
24 Literature and a reference Tarantola: Inverse problem theory, Methods for data fitting and model parameter estimation, Elsevier, 1987 Rodgers: Inverse Methods for Atmospheric Sounding: Theory and Practice, World Scientific, 2000 Menke: Geophysical data analysis: discrete inverse theory, Academic Press, 1984 Tamminen and Kyrölä, JGR, 106, 14377, 2001 Tamminen: Ph.D. thesis, FMI contributions 47, Day 3 Lecture 1 Retrieval techniques - Erkki Kyrölä 24
UV/VIS Limb Retrieval
UV/VIS Limb Retrieval Erkki Kyrölä Finnish Meteorological Institute 1. Data selection 2. Forward and inverse possibilities 3. Occultation: GOMOS inversion 4. Limb scattering: OSIRIS inversion 5. Summary
More informationMarkov 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 informationInverse problems and uncertainty quantification in remote sensing
1 / 38 Inverse problems and uncertainty quantification in remote sensing Johanna Tamminen Finnish Meterological Institute johanna.tamminen@fmi.fi ESA Earth Observation Summer School on Earth System Monitoring
More informationInverse Theory. COST WaVaCS Winterschool Venice, February Stefan Buehler Luleå University of Technology Kiruna
Inverse Theory COST WaVaCS Winterschool Venice, February 2011 Stefan Buehler Luleå University of Technology Kiruna Overview Inversion 1 The Inverse Problem 2 Simple Minded Approach (Matrix Inversion) 3
More informationConsider the joint probability, P(x,y), shown as the contours in the figure above. P(x) is given by the integral of P(x,y) over all values of y.
ATMO/OPTI 656b Spring 009 Bayesian Retrievals Note: This follows the discussion in Chapter of Rogers (000) As we have seen, the problem with the nadir viewing emission measurements is they do not contain
More informationAEROSOL MODEL SELECTION AND UNCERTAINTY MODELLING BY RJMCMC TECHNIQUE
AEROSOL MODEL SELECTION AND UNCERTAINTY MODELLING BY RJMCMC TECHNIQUE Marko Laine 1, Johanna Tamminen 1, Erkki Kyrölä 1, and Heikki Haario 2 1 Finnish Meteorological Institute, Helsinki, Finland 2 Lappeenranta
More informationDensity 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 information4. DATA ASSIMILATION FUNDAMENTALS
4. DATA ASSIMILATION FUNDAMENTALS... [the atmosphere] "is a chaotic system in which errors introduced into the system can grow with time... As a consequence, data assimilation is a struggle between chaotic
More informationSTA 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 informationStrong 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 informationLecture 3. G. Cowan. Lecture 3 page 1. Lectures on Statistical Data Analysis
Lecture 3 1 Probability (90 min.) Definition, Bayes theorem, probability densities and their properties, catalogue of pdfs, Monte Carlo 2 Statistical tests (90 min.) general concepts, test statistics,
More informationDetection ASTR ASTR509 Jasper Wall Fall term. William Sealey Gosset
ASTR509-14 Detection William Sealey Gosset 1876-1937 Best known for his Student s t-test, devised for handling small samples for quality control in brewing. To many in the statistical world "Student" was
More informationSpatial 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 informationMarkov 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 informationLecture 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 informationProbing the covariance matrix
Probing the covariance matrix Kenneth M. Hanson Los Alamos National Laboratory (ret.) BIE Users Group Meeting, September 24, 2013 This presentation available at http://kmh-lanl.hansonhub.com/ LA-UR-06-5241
More informationMCMC 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 informationBrandon C. Kelly (Harvard Smithsonian Center for Astrophysics)
Brandon C. Kelly (Harvard Smithsonian Center for Astrophysics) Probability quantifies randomness and uncertainty How do I estimate the normalization and logarithmic slope of a X ray continuum, assuming
More informationData Analysis I. Dr Martin Hendry, Dept of Physics and Astronomy University of Glasgow, UK. 10 lectures, beginning October 2006
Astronomical p( y x, I) p( x, I) p ( x y, I) = p( y, I) Data Analysis I Dr Martin Hendry, Dept of Physics and Astronomy University of Glasgow, UK 10 lectures, beginning October 2006 4. Monte Carlo Methods
More informationBayesian rules of probability as principles of logic [Cox] Notation: pr(x I) is the probability (or pdf) of x being true given information I
Bayesian rules of probability as principles of logic [Cox] Notation: pr(x I) is the probability (or pdf) of x being true given information I 1 Sum rule: If set {x i } is exhaustive and exclusive, pr(x
More informationShort tutorial on data assimilation
Mitglied der Helmholtz-Gemeinschaft Short tutorial on data assimilation 23 June 2015 Wolfgang Kurtz & Harrie-Jan Hendricks Franssen Institute of Bio- and Geosciences IBG-3 (Agrosphere), Forschungszentrum
More informationThe Kalman Filter ImPr Talk
The Kalman Filter ImPr Talk Ged Ridgway Centre for Medical Image Computing November, 2006 Outline What is the Kalman Filter? State Space Models Kalman Filter Overview Bayesian Updating of Estimates Kalman
More informationGaussian Process Approximations of Stochastic Differential Equations
Gaussian Process Approximations of Stochastic Differential Equations Cédric Archambeau Dan Cawford Manfred Opper John Shawe-Taylor May, 2006 1 Introduction Some of the most complex models routinely run
More informationVariational 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 informationMarkov chain Monte Carlo methods for high dimensional inversion in remote sensing
J. R. Statist. Soc. B (04) 66, Part 3, pp. 591 7 Markov chain Monte Carlo methods for high dimensional inversion in remote sensing H. Haario and M. Laine, University of Helsinki, Finland M. Lehtinen, University
More informationKalman filtering and friends: Inference in time series models. Herke van Hoof slides mostly by Michael Rubinstein
Kalman filtering and friends: Inference in time series models Herke van Hoof slides mostly by Michael Rubinstein Problem overview Goal Estimate most probable state at time k using measurement up to time
More informationBayesian Regression Linear and Logistic Regression
When we want more than point estimates Bayesian Regression Linear and Logistic Regression Nicole Beckage Ordinary Least Squares Regression and Lasso Regression return only point estimates But what if we
More informationLecture 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 informationSTA414/2104 Statistical Methods for Machine Learning II
STA414/2104 Statistical Methods for Machine Learning II Murat A. Erdogdu & David Duvenaud Department of Computer Science Department of Statistical Sciences Lecture 3 Slide credits: Russ Salakhutdinov Announcements
More informationLocal Positioning with Parallelepiped Moving Grid
Local Positioning with Parallelepiped Moving Grid, WPNC06 16.3.2006, niilo.sirola@tut.fi p. 1/?? TA M P E R E U N I V E R S I T Y O F T E C H N O L O G Y M a t h e m a t i c s Local Positioning with Parallelepiped
More informationAdvanced 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 informationGaussian Processes for Machine Learning
Gaussian Processes for Machine Learning Carl Edward Rasmussen Max Planck Institute for Biological Cybernetics Tübingen, Germany carl@tuebingen.mpg.de Carlos III, Madrid, May 2006 The actual science of
More informationPrediction of Data with help of the Gaussian Process Method
of Data with help of the Gaussian Process Method R. Preuss, U. von Toussaint Max-Planck-Institute for Plasma Physics EURATOM Association 878 Garching, Germany March, Abstract The simulation of plasma-wall
More informationLecture 5. G. Cowan Lectures on Statistical Data Analysis Lecture 5 page 1
Lecture 5 1 Probability (90 min.) Definition, Bayes theorem, probability densities and their properties, catalogue of pdfs, Monte Carlo 2 Statistical tests (90 min.) general concepts, test statistics,
More informationMultiple Scenario Inversion of Reflection Seismic Prestack Data
Downloaded from orbit.dtu.dk on: Nov 28, 2018 Multiple Scenario Inversion of Reflection Seismic Prestack Data Hansen, Thomas Mejer; Cordua, Knud Skou; Mosegaard, Klaus Publication date: 2013 Document Version
More informationLecture 7 and 8: Markov Chain Monte Carlo
Lecture 7 and 8: Markov Chain Monte Carlo 4F13: Machine Learning Zoubin Ghahramani and Carl Edward Rasmussen Department of Engineering University of Cambridge http://mlg.eng.cam.ac.uk/teaching/4f13/ Ghahramani
More informationParameter Estimation. William H. Jefferys University of Texas at Austin Parameter Estimation 7/26/05 1
Parameter Estimation William H. Jefferys University of Texas at Austin bill@bayesrules.net Parameter Estimation 7/26/05 1 Elements of Inference Inference problems contain two indispensable elements: Data
More informationBayesian inversion (I):
Bayesian inversion (I): Advanced topics Wojciech Dȩbski Instytut Geofizyki PAN debski@igf.edu.pl Wydział Fizyki UW, 13.10.2004 Wydział Fizyki UW Warszawa, 13.10.2004 (1) Plan of the talk I. INTRODUCTION
More informationGaussian Process Approximations of Stochastic Differential Equations
Gaussian Process Approximations of Stochastic Differential Equations Cédric Archambeau Centre for Computational Statistics and Machine Learning University College London c.archambeau@cs.ucl.ac.uk CSML
More informationDynamic System Identification using HDMR-Bayesian Technique
Dynamic System Identification using HDMR-Bayesian Technique *Shereena O A 1) and Dr. B N Rao 2) 1), 2) Department of Civil Engineering, IIT Madras, Chennai 600036, Tamil Nadu, India 1) ce14d020@smail.iitm.ac.in
More informationWinter 2019 Math 106 Topics in Applied Mathematics. Lecture 8: Importance Sampling
Winter 2019 Math 106 Topics in Applied Mathematics Data-driven Uncertainty Quantification Yoonsang Lee (yoonsang.lee@dartmouth.edu) Lecture 8: Importance Sampling 8.1 Importance Sampling Importance sampling
More informationA Bayesian Treatment of Linear Gaussian Regression
A Bayesian Treatment of Linear Gaussian Regression Frank Wood December 3, 2009 Bayesian Approach to Classical Linear Regression In classical linear regression we have the following model y β, σ 2, X N(Xβ,
More informationIntroduction 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 informationProbabilistic 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 informationBayesian Inference in Astronomy & Astrophysics A Short Course
Bayesian Inference in Astronomy & Astrophysics A Short Course Tom Loredo Dept. of Astronomy, Cornell University p.1/37 Five Lectures Overview of Bayesian Inference From Gaussians to Periodograms Learning
More informationInverse Problems in the Bayesian Framework
Inverse Problems in the Bayesian Framework Daniela Calvetti Case Western Reserve University Cleveland, Ohio Raleigh, NC, July 2016 Bayes Formula Stochastic model: Two random variables X R n, B R m, where
More informationApproximate Bayesian computation: an application to weak-lensing peak counts
STATISTICAL CHALLENGES IN MODERN ASTRONOMY VI Approximate Bayesian computation: an application to weak-lensing peak counts Chieh-An Lin & Martin Kilbinger SAp, CEA Saclay Carnegie Mellon University, Pittsburgh
More informationFundamentals of Data Assimilation
National Center for Atmospheric Research, Boulder, CO USA GSI Data Assimilation Tutorial - June 28-30, 2010 Acknowledgments and References WRFDA Overview (WRF Tutorial Lectures, H. Huang and D. Barker)
More informationBayesian analysis in nuclear physics
Bayesian analysis in nuclear physics Ken Hanson T-16, Nuclear Physics; Theoretical Division Los Alamos National Laboratory Tutorials presented at LANSCE Los Alamos Neutron Scattering Center July 25 August
More informationCovariance 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 informationA new Hierarchical Bayes approach to ensemble-variational data assimilation
A new Hierarchical Bayes approach to ensemble-variational data assimilation Michael Tsyrulnikov and Alexander Rakitko HydroMetCenter of Russia College Park, 20 Oct 2014 Michael Tsyrulnikov and Alexander
More informationMCMC 2: Lecture 3 SIR models - more topics. Phil O Neill Theo Kypraios School of Mathematical Sciences University of Nottingham
MCMC 2: Lecture 3 SIR models - more topics Phil O Neill Theo Kypraios School of Mathematical Sciences University of Nottingham Contents 1. What can be estimated? 2. Reparameterisation 3. Marginalisation
More informationChapter 3: Maximum-Likelihood & Bayesian Parameter Estimation (part 1)
HW 1 due today Parameter Estimation Biometrics CSE 190 Lecture 7 Today s lecture was on the blackboard. These slides are an alternative presentation of the material. CSE190, Winter10 CSE190, Winter10 Chapter
More informationISTINA - : Investigation of Sensitivity Tendencies and Inverse Numerical Algorithm advances in aerosol remote sensing
STNA - : nvestigation of Sensitivity Tendencies and nverse Numerical Algorithm advances in aerosol remote sensing B. Torres, O. Dubovik, D. Fuertes, and P. Litvinov GRASP- SAS, LOA, Universite Lille-1,
More informationApril 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 informationModeling and state estimation Examples State estimation Probabilities Bayes filter Particle filter. Modeling. CSC752 Autonomous Robotic Systems
Modeling CSC752 Autonomous Robotic Systems Ubbo Visser Department of Computer Science University of Miami February 21, 2017 Outline 1 Modeling and state estimation 2 Examples 3 State estimation 4 Probabilities
More informationParametric Models. Dr. Shuang LIANG. School of Software Engineering TongJi University Fall, 2012
Parametric Models Dr. Shuang LIANG School of Software Engineering TongJi University Fall, 2012 Today s Topics Maximum Likelihood Estimation Bayesian Density Estimation Today s Topics Maximum Likelihood
More informationMAT Inverse Problems, Part 2: Statistical Inversion
MAT-62006 Inverse Problems, Part 2: Statistical Inversion S. Pursiainen Department of Mathematics, Tampere University of Technology Spring 2015 Overview The statistical inversion approach is based on the
More informationBayesian 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 informationConstrained data assimilation. W. Carlisle Thacker Atlantic Oceanographic and Meteorological Laboratory Miami, Florida USA
Constrained data assimilation W. Carlisle Thacker Atlantic Oceanographic and Meteorological Laboratory Miami, Florida 33149 USA Plan Range constraints: : HYCOM layers have minimum thickness. Optimal interpolation:
More informationEfficient 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 informationFrequentist-Bayesian Model Comparisons: A Simple Example
Frequentist-Bayesian Model Comparisons: A Simple Example Consider data that consist of a signal y with additive noise: Data vector (N elements): D = y + n The additive noise n has zero mean and diagonal
More informationLecture: Gaussian Process Regression. STAT 6474 Instructor: Hongxiao Zhu
Lecture: Gaussian Process Regression STAT 6474 Instructor: Hongxiao Zhu Motivation Reference: Marc Deisenroth s tutorial on Robot Learning. 2 Fast Learning for Autonomous Robots with Gaussian Processes
More informationRecent Advances in Bayesian Inference for Inverse Problems
Recent Advances in Bayesian Inference for Inverse Problems Felix Lucka University College London, UK f.lucka@ucl.ac.uk Applied Inverse Problems Helsinki, May 25, 2015 Bayesian Inference for Inverse Problems
More informationFundamental Probability and Statistics
Fundamental Probability and Statistics "There are known knowns. These are things we know that we know. There are known unknowns. That is to say, there are things that we know we don't know. But there are
More informationLecture notes on Regression: Markov Chain Monte Carlo (MCMC)
Lecture notes on Regression: Markov Chain Monte Carlo (MCMC) Dr. Veselina Kalinova, Max Planck Institute for Radioastronomy, Bonn Machine Learning course: the elegant way to extract information from data,
More informationErgodicity in data assimilation methods
Ergodicity in data assimilation methods David Kelly Andy Majda Xin Tong Courant Institute New York University New York NY www.dtbkelly.com April 15, 2016 ETH Zurich David Kelly (CIMS) Data assimilation
More informationStatistical Data Analysis Stat 3: p-values, parameter estimation
Statistical Data Analysis Stat 3: p-values, parameter estimation London Postgraduate Lectures on Particle Physics; University of London MSci course PH4515 Glen Cowan Physics Department Royal Holloway,
More informationUniversität Potsdam Institut für Informatik Lehrstuhl Maschinelles Lernen. Bayesian Learning. Tobias Scheffer, Niels Landwehr
Universität Potsdam Institut für Informatik Lehrstuhl Maschinelles Lernen Bayesian Learning Tobias Scheffer, Niels Landwehr Remember: Normal Distribution Distribution over x. Density function with parameters
More informationCOMP 551 Applied Machine Learning Lecture 20: Gaussian processes
COMP 55 Applied Machine Learning Lecture 2: Gaussian processes Instructor: Ryan Lowe (ryan.lowe@cs.mcgill.ca) Slides mostly by: (herke.vanhoof@mcgill.ca) Class web page: www.cs.mcgill.ca/~hvanho2/comp55
More informationProbability and Estimation. Alan Moses
Probability and Estimation Alan Moses Random variables and probability A random variable is like a variable in algebra (e.g., y=e x ), but where at least part of the variability is taken to be stochastic.
More information1 Bayesian Linear Regression (BLR)
Statistical Techniques in Robotics (STR, S15) Lecture#10 (Wednesday, February 11) Lecturer: Byron Boots Gaussian Properties, Bayesian Linear Regression 1 Bayesian Linear Regression (BLR) In linear regression,
More informationMore Information! Conditional Probability and Inverse Methods Peter Zoogman Harvard Atmospheres Journal Club April 12, 2012
More Information! Conditional Probability and Inverse Methods Peter Zoogman Harvard Atmospheres Journal Club April 12, 2012 Bayesian Analysis not very intuitive! Humans are bad at combining a priori information
More informationJanne Hakkarainen ON STATE AND PARAMETER ESTIMATION IN CHAOTIC SYSTEMS. Acta Universitatis Lappeenrantaensis 545
Janne Hakkarainen ON STATE AND PARAMETER ESTIMATION IN CHAOTIC SYSTEMS Thesis for the degree of Doctor of Philosophy to be presented with due permission for public examination and criticism in the Auditorium
More informationAircraft Validation of Infrared Emissivity derived from Advanced InfraRed Sounder Satellite Observations
Aircraft Validation of Infrared Emissivity derived from Advanced InfraRed Sounder Satellite Observations Robert Knuteson, Fred Best, Steve Dutcher, Ray Garcia, Chris Moeller, Szu Chia Moeller, Henry Revercomb,
More informationBayesian Inference for Wind Field Retrieval
Bayesian Inference for Wind Field Retrieval Ian T. Nabney 1, Dan Cornford Neural Computing Research Group, Aston University, Aston Triangle, Birmingham B4 7ET, UK Christopher K. I. Williams Division of
More informationIntroduction to Systems Analysis and Decision Making Prepared by: Jakub Tomczak
Introduction to Systems Analysis and Decision Making Prepared by: Jakub Tomczak 1 Introduction. Random variables During the course we are interested in reasoning about considered phenomenon. In other words,
More informationIntroduc)on to Bayesian Methods
Introduc)on to Bayesian Methods Bayes Rule py x)px) = px! y) = px y)py) py x) = px y)py) px) px) =! px! y) = px y)py) y py x) = py x) =! y "! y px y)py) px y)py) px y)py) px y)py)dy Bayes Rule py x) =
More informationWinter 2019 Math 106 Topics in Applied Mathematics. Lecture 1: Introduction
Winter 2019 Math 106 Topics in Applied Mathematics Data-driven Uncertainty Quantification Yoonsang Lee (yoonsang.lee@dartmouth.edu) Lecture 1: Introduction 19 Winter M106 Class: MWF 12:50-1:55 pm @ 200
More informationPoint 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 informationLearning with Noisy Labels. Kate Niehaus Reading group 11-Feb-2014
Learning with Noisy Labels Kate Niehaus Reading group 11-Feb-2014 Outline Motivations Generative model approach: Lawrence, N. & Scho lkopf, B. Estimating a Kernel Fisher Discriminant in the Presence of
More informationProbabilistic Machine Learning
Probabilistic Machine Learning Bayesian Nets, MCMC, and more Marek Petrik 4/18/2017 Based on: P. Murphy, K. (2012). Machine Learning: A Probabilistic Perspective. Chapter 10. Conditional Independence Independent
More informationLeïla Haegel University of Geneva
Picture found at: http://www.ps.uci.edu/~tomba/sk/tscan/pictures.html Leïla Haegel University of Geneva Seminar at the University of Sheffield April 27th, 2017 1 Neutrinos are: the only particle known
More informationBlind Equalization via Particle Filtering
Blind Equalization via Particle Filtering Yuki Yoshida, Kazunori Hayashi, Hideaki Sakai Department of System Science, Graduate School of Informatics, Kyoto University Historical Remarks A sequential Monte
More informationGOMOS data characterisation and error estimation
Atmos. Chem. Phys.,, 95 959, www.atmos-chem-phys.net//95// doi:.594/acp--95- Author(s). CC Attribution 3. License. Atmospheric Chemistry and Physics GOMOS data characterisation and error estimation J.
More informationSTA 4273H: Statistical Machine Learning
STA 4273H: Statistical Machine Learning Russ Salakhutdinov Department of Statistics! rsalakhu@utstat.toronto.edu! http://www.utstat.utoronto.ca/~rsalakhu/ Sidney Smith Hall, Room 6002 Lecture 3 Linear
More informationParameter 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 informationTrans-dimensional inverse problems, model comparison and the evidence
Geophys. J. Int. (2006) 67, 528 542 doi: 0./j.365-246X.2006.0355.x GJI Geodesy, potential field and applied geophysics Trans-dimensional inverse problems, model comparison and the evidence M. Sambridge,
More informationConditions for successful data assimilation
Conditions for successful data assimilation Matthias Morzfeld *,**, Alexandre J. Chorin *,**, Peter Bickel # * Department of Mathematics University of California, Berkeley ** Lawrence Berkeley National
More informationCS 630 Basic Probability and Information Theory. Tim Campbell
CS 630 Basic Probability and Information Theory Tim Campbell 21 January 2003 Probability Theory Probability Theory is the study of how best to predict outcomes of events. An experiment (or trial or event)
More informationLecture : Probabilistic Machine Learning
Lecture : Probabilistic Machine Learning Riashat Islam Reasoning and Learning Lab McGill University September 11, 2018 ML : Many Methods with Many Links Modelling Views of Machine Learning Machine Learning
More informationNonparametric 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 informationUsing Bayesian Analysis to Study Tokamak Plasma Equilibrium
Using Bayesian Analysis to Study Tokamak Plasma Equilibrium Greg T. von Nessi (joint work with Matthew Hole, Jakob Svensson, Lynton Appel, Boyd Blackwell, John Howard, David Pretty and Jason Bertram) Outline
More informationCourse on Inverse Problems
California Institute of Technology Division of Geological and Planetary Sciences March 26 - May 25, 2007 Course on Inverse Problems Albert Tarantola Institut de Physique du Globe de Paris Lesson XVI CONCLUSION
More informationLikelihood NIPS July 30, Gaussian Process Regression with Student-t. Likelihood. Jarno Vanhatalo, Pasi Jylanki and Aki Vehtari NIPS-2009
with with July 30, 2010 with 1 2 3 Representation Representation for Distribution Inference for the Augmented Model 4 Approximate Laplacian Approximation Introduction to Laplacian Approximation Laplacian
More informationPhysics 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 informationFundamentals of Data Assimila1on
014 GSI Community Tutorial NCAR Foothills Campus, Boulder, CO July 14-16, 014 Fundamentals of Data Assimila1on Milija Zupanski Cooperative Institute for Research in the Atmosphere Colorado State University
More informationMachine Learning. Lecture 4: Regularization and Bayesian Statistics. Feng Li. https://funglee.github.io
Machine Learning Lecture 4: Regularization and Bayesian Statistics Feng Li fli@sdu.edu.cn https://funglee.github.io School of Computer Science and Technology Shandong University Fall 207 Overfitting Problem
More informationStatistical Methods in Particle Physics Lecture 1: Bayesian methods
Statistical Methods in Particle Physics Lecture 1: Bayesian methods SUSSP65 St Andrews 16 29 August 2009 Glen Cowan Physics Department Royal Holloway, University of London g.cowan@rhul.ac.uk www.pp.rhul.ac.uk/~cowan
More informationCourse on Inverse Problems Albert Tarantola
California Institute of Technology Division of Geological and Planetary Sciences March 26 - May 25, 27 Course on Inverse Problems Albert Tarantola Institut de Physique du Globe de Paris CONCLUSION OF THE
More information