SlideShare a Scribd company logo
Bayesian Inference and Uncertainty Quantification for Inverse Problems
Matt Moores
@MooresMt
https://uow.edu.au/~mmoores/
Centre for Environmental Informatics (CEI), University of Wollongong, NSW, Australia
TIDE Hub Seminar Series
joint with Matthew Berry, Mark Nelson, Brian Monaghan and Raymond Longbottom
1 / 19
Introduction
Physical model F

t; ~
θ

Unknown parameters ~
θ = {θ1, . . . , θk}
Initial conditions F0 and derivatives ∂F
∂t , etc.
2 / 19
Introduction
Physical model F

t; ~
θ

Unknown parameters ~
θ = {θ1, . . . , θk}
Initial conditions F0 and derivatives ∂F
∂t , etc.
Observed data yt at times t = 1, . . . , T
yt = F

t; ~
θ

+ εt
where εt is random noise.
2 / 19
Example
Growth curve:
dF
dt
= −βγt
log γ
3 / 19
Example
Growth curve:
dF
dt
= −βγt
log γ
Solution:
F

ti; ~
θ

= α − βγti
at time points t1, . . . , tn with initial condition F0 = α − β.
3 / 19
Example
Growth curve:
dF
dt
= −βγt
log γ
Solution:
F

ti; ~
θ

= α − βγti
at time points t1, . . . , tn with initial condition F0 = α − β.
Noisy observations:
yi = F(ti; α, β, γ) + εi
where εi ∼ N(0, σ2
) is additive Gaussian noise.
Unknown parameters ~
θ = {α, β, γ, σ2
}
3 / 19
Observed Data for n = 27 dugongs
4 / 19
Nonlinear Least Squares
nlm - nls(y ~ alpha - beta * gammˆt, data=dat,
start=list(alpha=1, beta=1, gamm=0.9))
summary(nlm)
##
## Formula: y ~ alpha - beta * gamm^t
##
## Parameters:
## Estimate Std. Error t value Pr(|t|)
## alpha 2.65807 0.06151 43.21  2e-16 ***
## beta 0.96352 0.06968 13.83 6.3e-13 ***
## gamm 0.87146 0.02460 35.42  2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.09525 on 24 degrees of freedom
5 / 19
Drawbacks
No guarantee of convexity:
will only find the nearest local optimum
results are completely dependent on initialization
6 / 19
Drawbacks
No guarantee of convexity:
will only find the nearest local optimum
results are completely dependent on initialization
Standard errors are underestimated:
95% profile CI for α and β only achieve ~83% coverage
6 / 19
Inverse Problem
Find parameters ~
θ consistent with ~
y
7 / 19
Inverse Problem
Find parameters ~
θ consistent with ~
y
Physical model F : Θ → Y doesn’t need to be invertible:
although lack of identifiability for ~
θ can create problems
we don’t need F in closed form — can use numeric solver for DE
7 / 19
Inverse Problem
Find parameters ~
θ consistent with ~
y
Physical model F : Θ → Y doesn’t need to be invertible:
although lack of identifiability for ~
θ can create problems
we don’t need F in closed form — can use numeric solver for DE
Likelihood L(~
y | ~
θ) is based on
E[Yi | ~
θ] = F

ti; ~
θ

as well as distribution of random noise (not necessarily Gaussian)
7 / 19
Bayesian Inference
We assign prior distributions to the parameters ~
θ:
π(log{α}) ∼ N(0, 100)
π(log{β}) ∼ N(0, 100)
π(logit{γ}) ∼ N(0, 100)
π

1
σ2

∼ Ga
ν0
2
,
2
ν0s2
0
!
8 / 19
Astfalck, Cripps, Gosling, Hodkiewicz  Milne (2018) Ocean Engineering 161: 268–276.
Bayesian Inference
We assign prior distributions to the parameters ~
θ:
π(log{α}) ∼ N(0, 100)
π(log{β}) ∼ N(0, 100)
π(logit{γ}) ∼ N(0, 100)
π

1
σ2

∼ Ga
ν0
2
,
2
ν0s2
0
!
Then the joint posterior distribution is:
π

~
θ | ~
y

=
L(~
y | ~
θ)π(~
θ)
π(~
y)
where the normalising constant π(~
y) =
R ∞
0
R ∞
0
R ∞
0
R 1
0 L(~
y | α, β, γ, σ) dπ(γ)dπ(β)dπ(α)dπ(σ)
is intractable.
8 / 19
Astfalck, Cripps, Gosling, Hodkiewicz  Milne (2018) Ocean Engineering 161: 268–276.
Markov Chain Monte Carlo
Algorithm 1 Random-walk Metropolis sampler for π

~
θ | ~
y

1: Initialize α0 ∼ π(log{α}), β0 ∼ π(log{β}), γ0 ∼ π(logit{γ}), σ0 ∼ π

1
σ2

2: Solve ~
µ0 = ~
F

~
t; α0, β0, γ0

and calculate `0 = log

L(~
y | ~
µ0, σ2
0)
	
3: for iterations j = 1, . . . , J do
4: Propose log{α∗} ∼ N(log{αj−1}, σ2
α), log{β∗} ∼ N(log{βj−1}, σ2
β), logit{γ∗} ∼
N(logit{γj−1}, σ2
γ)
5: Solve ~
µ∗ = ~
F

~
t; α∗, β∗, γ∗

6: Propose σ∗ ∼ π

1
σ2 | ~
µ∗,~
y

and calculate `∗ = log

L(~
y | ~
µ∗, σ2
∗)
	
7: Calculate
ρt =
exp{`∗}π(~
θ∗)q(~
θj−1 | ~
θ∗)
exp{`j−1}π(~
θj−1)q(~
θ∗ | ~
θj−1)
8: Accept ~
θj = ~
θ∗ with probability min{ρt, 1} else ~
θj = ~
θj−1
9: end for 9 / 19
Markov Chains
10 / 19
Posterior Distributions
11 / 19
95% Predictive Interval
12 / 19
Thermogravimetric Analysis
The TGA model for a single reaction involves the Arrhenius equations:
dM
dt
= −MA exp

−
E
RT

dT
dt
= α
where M is mass fraction, T is temperature, R is the ideal gas constant. The initial mass M0,
the initial temperature T0, and the heating rate α are experimentally controlled.
The unknown parameters ~
θ are
A pre-exponential factor
E activation energy
σ2 variance of the additive Gaussian noise
13 / 19
Reparameterisation
We have good prior information for E (physically interpretable) and σ2 (measurement noise),
but not for A.
The distributions for A and E are also very highly correlated, slowing the mixing of MCMC.
14 / 19
Reparameterisation
We have good prior information for E (physically interpretable) and σ2 (measurement noise),
but not for A.
The distributions for A and E are also very highly correlated, slowing the mixing of MCMC.
Instead, we reparameterise the model in terms of the temperature, Tm, at which the rate dM
dt is
maximised:
A exp

−
E
RTm

=
E
RT2
m
α
In our MCMC algorithm, we first propose q(E∗ | Ej−1), then propose q(Tm∗ | Tm,j−1). The
value of A∗ can then be obtained from the equation above.
We solve ~
F

~
t; E∗, A∗

numerically using a Runge-Kutta method.
14 / 19
Posterior Distributions
634.6 634.7 634.8 634.9 635.0 635.1 635.2 635.3
Tm
0
1
2
3
4
5
84750 85000 85250 85500 85750 86000 86250 86500 86750
E
0.0000
0.0002
0.0004
0.0006
0.0008
0.0010
0.0012
0.0014
0.0016
0.08 0.10 0.12 0.14
0
10
20
30
40
50
2400000 2600000 2800000 3000000 3200000 3400000 3600000
A
0.0000000
0.0000005
0.0000010
0.0000015
0.0000020
0.0000025
15 / 19
Posterior for Functions of ~
θ
A function of a random variable is also a random variable.
Now that we have random samples from our posterior π(A, E, Tm, σ2 | ~
y), we can use our
model to obtain predictions for any measurable function of the parameters, g(A, E, Tm).
16 / 19
Posterior for Functions of ~
θ
A function of a random variable is also a random variable.
Now that we have random samples from our posterior π(A, E, Tm, σ2 | ~
y), we can use our
model to obtain predictions for any measurable function of the parameters, g(A, E, Tm).
For example, the critical length of a stockpile:
Lcr = g(A, E, Tm) = K
v
u
u
texp
n
E
RT2
m
o
A
RT2
m
E
E[g(A, E, Tm) | ~
y] =
Z 1000
0
Z ∞
0
Z ∞
0
g(A, E, Tm) dπ(A, E, Tm | ~
y)
≈
J
X
j=1
g(Aj, Ej, Tj)
where K is a constant and {Aj, Ej, Tj}J
j=1 are the MCMC samples (after discarding burn-in).
16 / 19
Posterior for Lcr
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
17 / 19
Advanced Methods
Parallel and distributed computation
Sequential updating of π(~
θ | ~
y) and streaming inference
Spatio-temporal modelling of F(x, y, z, t; ~
θ)
Emulation of the forward map Θ → Y
(e.g. using artificial neural networks)
Approximating the likelihood L (e.g. ABC)
Accounting for multi-modality
(e.g. ill-posed problems)
Accounting for error in the numerical solver (probabilistic numerics)
Accounting for model misspecification (discrepancy function)
Multi-level Monte Carlo methods
Hamiltonian Monte Carlo 18 / 19
Further Reading
R. J. Longbottom, B. J. Monaghan, D. J. Pinson, N. A. S. Webster  S. J. Chew
In situ Phase Analysis during Self-sintering of BOS Filter Cake for Improved Recycling.
ISIJ International 60(11): 2436–2445, 2020.
A. M. Stuart
Inverse problems: A Bayesian perspective.
Acta Numerica 19: 451—559, 2010.
L.C. Astfalck, E.J. Cripps, J.P. Gosling, M.R. Hodkiewicz  I.A. Milne
Expert elicitation of directional metocean parameters. Ocean Engineering 161: 268–276, 2018.
B. P. Carlin  A. E. Gelfand
An iterative Monte Carlo method for nonconjugate Bayesian analysis.
Statistics  Computing 1(2): 119–128, 1991.
19 / 19

More Related Content

What's hot

SPU Optimizations-part 1
SPU Optimizations-part 1SPU Optimizations-part 1
SPU Optimizations-part 1Naughty Dog
 
A Note on Latent LSTM Allocation
A Note on Latent LSTM AllocationA Note on Latent LSTM Allocation
A Note on Latent LSTM AllocationTomonari Masada
 
The all-electron GW method based on WIEN2k: Implementation and applications.
The all-electron GW method based on WIEN2k: Implementation and applications.The all-electron GW method based on WIEN2k: Implementation and applications.
The all-electron GW method based on WIEN2k: Implementation and applications.ABDERRAHMANE REGGAD
 
Multilayer Neural Networks
Multilayer Neural NetworksMultilayer Neural Networks
Multilayer Neural NetworksESCOM
 
A nonlinear approximation of the Bayesian Update formula
A nonlinear approximation of the Bayesian Update formulaA nonlinear approximation of the Bayesian Update formula
A nonlinear approximation of the Bayesian Update formulaAlexander Litvinenko
 
Random Number Generators 2018
Random Number Generators 2018Random Number Generators 2018
Random Number Generators 2018rinnocente
 
Convex Optimization Modelling with CVXOPT
Convex Optimization Modelling with CVXOPTConvex Optimization Modelling with CVXOPT
Convex Optimization Modelling with CVXOPTandrewmart11
 
A study of the worst case ratio of a simple algorithm for simple assembly lin...
A study of the worst case ratio of a simple algorithm for simple assembly lin...A study of the worst case ratio of a simple algorithm for simple assembly lin...
A study of the worst case ratio of a simple algorithm for simple assembly lin...narmo
 
ADAPTIVE FUZZY KERNEL CLUSTERING ALGORITHM
ADAPTIVE FUZZY KERNEL CLUSTERING ALGORITHMADAPTIVE FUZZY KERNEL CLUSTERING ALGORITHM
ADAPTIVE FUZZY KERNEL CLUSTERING ALGORITHMijfls
 
Aaex4 group2(中英夾雜)
Aaex4 group2(中英夾雜)Aaex4 group2(中英夾雜)
Aaex4 group2(中英夾雜)Shiang-Yun Yang
 
Hyperparameter optimization with approximate gradient
Hyperparameter optimization with approximate gradientHyperparameter optimization with approximate gradient
Hyperparameter optimization with approximate gradientFabian Pedregosa
 
Aaex7 group2(中英夾雜)
Aaex7 group2(中英夾雜)Aaex7 group2(中英夾雜)
Aaex7 group2(中英夾雜)Shiang-Yun Yang
 
Linear Cryptanalysis Lecture 線形解読法
Linear Cryptanalysis Lecture 線形解読法Linear Cryptanalysis Lecture 線形解読法
Linear Cryptanalysis Lecture 線形解読法Kai Katsumata
 

What's hot (18)

Graph Kernelpdf
Graph KernelpdfGraph Kernelpdf
Graph Kernelpdf
 
SPU Optimizations-part 1
SPU Optimizations-part 1SPU Optimizations-part 1
SPU Optimizations-part 1
 
A Note on Latent LSTM Allocation
A Note on Latent LSTM AllocationA Note on Latent LSTM Allocation
A Note on Latent LSTM Allocation
 
Feedback Vertex Set
Feedback Vertex SetFeedback Vertex Set
Feedback Vertex Set
 
The all-electron GW method based on WIEN2k: Implementation and applications.
The all-electron GW method based on WIEN2k: Implementation and applications.The all-electron GW method based on WIEN2k: Implementation and applications.
The all-electron GW method based on WIEN2k: Implementation and applications.
 
Multilayer Neural Networks
Multilayer Neural NetworksMultilayer Neural Networks
Multilayer Neural Networks
 
Lecture9 xing
Lecture9 xingLecture9 xing
Lecture9 xing
 
A nonlinear approximation of the Bayesian Update formula
A nonlinear approximation of the Bayesian Update formulaA nonlinear approximation of the Bayesian Update formula
A nonlinear approximation of the Bayesian Update formula
 
Random Number Generators 2018
Random Number Generators 2018Random Number Generators 2018
Random Number Generators 2018
 
Convex Optimization Modelling with CVXOPT
Convex Optimization Modelling with CVXOPTConvex Optimization Modelling with CVXOPT
Convex Optimization Modelling with CVXOPT
 
A study of the worst case ratio of a simple algorithm for simple assembly lin...
A study of the worst case ratio of a simple algorithm for simple assembly lin...A study of the worst case ratio of a simple algorithm for simple assembly lin...
A study of the worst case ratio of a simple algorithm for simple assembly lin...
 
ADAPTIVE FUZZY KERNEL CLUSTERING ALGORITHM
ADAPTIVE FUZZY KERNEL CLUSTERING ALGORITHMADAPTIVE FUZZY KERNEL CLUSTERING ALGORITHM
ADAPTIVE FUZZY KERNEL CLUSTERING ALGORITHM
 
Aaex4 group2(中英夾雜)
Aaex4 group2(中英夾雜)Aaex4 group2(中英夾雜)
Aaex4 group2(中英夾雜)
 
Hyperparameter optimization with approximate gradient
Hyperparameter optimization with approximate gradientHyperparameter optimization with approximate gradient
Hyperparameter optimization with approximate gradient
 
Aaex7 group2(中英夾雜)
Aaex7 group2(中英夾雜)Aaex7 group2(中英夾雜)
Aaex7 group2(中英夾雜)
 
Subquad multi ff
Subquad multi ffSubquad multi ff
Subquad multi ff
 
Linear Cryptanalysis Lecture 線形解読法
Linear Cryptanalysis Lecture 線形解読法Linear Cryptanalysis Lecture 線形解読法
Linear Cryptanalysis Lecture 線形解読法
 
Minimum spanning tree
Minimum spanning treeMinimum spanning tree
Minimum spanning tree
 

Similar to Bayesian Inference and Uncertainty Quantification for Inverse Problems

Exact Sum Rules for Vector Channel at Finite Temperature and its Applications...
Exact Sum Rules for Vector Channel at Finite Temperature and its Applications...Exact Sum Rules for Vector Channel at Finite Temperature and its Applications...
Exact Sum Rules for Vector Channel at Finite Temperature and its Applications...Daisuke Satow
 
Generalized Nonlinear Models in R
Generalized Nonlinear Models in RGeneralized Nonlinear Models in R
Generalized Nonlinear Models in Rhtstatistics
 
Estimation of the score vector and observed information matrix in intractable...
Estimation of the score vector and observed information matrix in intractable...Estimation of the score vector and observed information matrix in intractable...
Estimation of the score vector and observed information matrix in intractable...Pierre Jacob
 
Computation of electromagnetic fields scattered from dielectric objects of un...
Computation of electromagnetic fields scattered from dielectric objects of un...Computation of electromagnetic fields scattered from dielectric objects of un...
Computation of electromagnetic fields scattered from dielectric objects of un...Alexander Litvinenko
 
Linear response theory and TDDFT
Linear response theory and TDDFT Linear response theory and TDDFT
Linear response theory and TDDFT Claudio Attaccalite
 
Finite-difference modeling, accuracy, and boundary conditions- Arthur Weglein...
Finite-difference modeling, accuracy, and boundary conditions- Arthur Weglein...Finite-difference modeling, accuracy, and boundary conditions- Arthur Weglein...
Finite-difference modeling, accuracy, and boundary conditions- Arthur Weglein...Arthur Weglein
 
Spectral sum rules for conformal field theories
Spectral sum rules for conformal field theoriesSpectral sum rules for conformal field theories
Spectral sum rules for conformal field theoriesSubham Dutta Chowdhury
 
chap4_lec1.ppt Engineering and technical
chap4_lec1.ppt Engineering and technicalchap4_lec1.ppt Engineering and technical
chap4_lec1.ppt Engineering and technicalshreenathji26
 
Estimation of the score vector and observed information matrix in intractable...
Estimation of the score vector and observed information matrix in intractable...Estimation of the score vector and observed information matrix in intractable...
Estimation of the score vector and observed information matrix in intractable...Pierre Jacob
 
Talk at SciCADE2013 about "Accelerated Multiple Precision ODE solver base on ...
Talk at SciCADE2013 about "Accelerated Multiple Precision ODE solver base on ...Talk at SciCADE2013 about "Accelerated Multiple Precision ODE solver base on ...
Talk at SciCADE2013 about "Accelerated Multiple Precision ODE solver base on ...Shizuoka Inst. Science and Tech.
 
Computation of Electromagnetic Fields Scattered from Dielectric Objects of Un...
Computation of Electromagnetic Fields Scattered from Dielectric Objects of Un...Computation of Electromagnetic Fields Scattered from Dielectric Objects of Un...
Computation of Electromagnetic Fields Scattered from Dielectric Objects of Un...Alexander Litvinenko
 

Similar to Bayesian Inference and Uncertainty Quantification for Inverse Problems (20)

Finite frequency H∞ control for wind turbine systems in T-S form
Finite frequency H∞ control for wind turbine systems in T-S formFinite frequency H∞ control for wind turbine systems in T-S form
Finite frequency H∞ control for wind turbine systems in T-S form
 
Ece4510 notes09
Ece4510 notes09Ece4510 notes09
Ece4510 notes09
 
Exact Sum Rules for Vector Channel at Finite Temperature and its Applications...
Exact Sum Rules for Vector Channel at Finite Temperature and its Applications...Exact Sum Rules for Vector Channel at Finite Temperature and its Applications...
Exact Sum Rules for Vector Channel at Finite Temperature and its Applications...
 
Generalized Nonlinear Models in R
Generalized Nonlinear Models in RGeneralized Nonlinear Models in R
Generalized Nonlinear Models in R
 
Estimation of the score vector and observed information matrix in intractable...
Estimation of the score vector and observed information matrix in intractable...Estimation of the score vector and observed information matrix in intractable...
Estimation of the score vector and observed information matrix in intractable...
 
Computation of electromagnetic fields scattered from dielectric objects of un...
Computation of electromagnetic fields scattered from dielectric objects of un...Computation of electromagnetic fields scattered from dielectric objects of un...
Computation of electromagnetic fields scattered from dielectric objects of un...
 
Poster Icqc
Poster IcqcPoster Icqc
Poster Icqc
 
CLIM Transition Workshop - Semiparametric Models for Extremes - Surya Tokdar,...
CLIM Transition Workshop - Semiparametric Models for Extremes - Surya Tokdar,...CLIM Transition Workshop - Semiparametric Models for Extremes - Surya Tokdar,...
CLIM Transition Workshop - Semiparametric Models for Extremes - Surya Tokdar,...
 
Linear response theory and TDDFT
Linear response theory and TDDFT Linear response theory and TDDFT
Linear response theory and TDDFT
 
Chang etal 2012a
Chang etal 2012aChang etal 2012a
Chang etal 2012a
 
Finite-difference modeling, accuracy, and boundary conditions- Arthur Weglein...
Finite-difference modeling, accuracy, and boundary conditions- Arthur Weglein...Finite-difference modeling, accuracy, and boundary conditions- Arthur Weglein...
Finite-difference modeling, accuracy, and boundary conditions- Arthur Weglein...
 
Spectral sum rules for conformal field theories
Spectral sum rules for conformal field theoriesSpectral sum rules for conformal field theories
Spectral sum rules for conformal field theories
 
chap4_lec1.ppt Engineering and technical
chap4_lec1.ppt Engineering and technicalchap4_lec1.ppt Engineering and technical
chap4_lec1.ppt Engineering and technical
 
Estimation of the score vector and observed information matrix in intractable...
Estimation of the score vector and observed information matrix in intractable...Estimation of the score vector and observed information matrix in intractable...
Estimation of the score vector and observed information matrix in intractable...
 
ACS 22LIE12 lab Manul.docx
ACS 22LIE12 lab Manul.docxACS 22LIE12 lab Manul.docx
ACS 22LIE12 lab Manul.docx
 
計算材料学
計算材料学計算材料学
計算材料学
 
A-tutorial.pdf
A-tutorial.pdfA-tutorial.pdf
A-tutorial.pdf
 
PCA on graph/network
PCA on graph/networkPCA on graph/network
PCA on graph/network
 
Talk at SciCADE2013 about "Accelerated Multiple Precision ODE solver base on ...
Talk at SciCADE2013 about "Accelerated Multiple Precision ODE solver base on ...Talk at SciCADE2013 about "Accelerated Multiple Precision ODE solver base on ...
Talk at SciCADE2013 about "Accelerated Multiple Precision ODE solver base on ...
 
Computation of Electromagnetic Fields Scattered from Dielectric Objects of Un...
Computation of Electromagnetic Fields Scattered from Dielectric Objects of Un...Computation of Electromagnetic Fields Scattered from Dielectric Objects of Un...
Computation of Electromagnetic Fields Scattered from Dielectric Objects of Un...
 

More from Matt Moores

bayesImageS: an R package for Bayesian image analysis
bayesImageS: an R package for Bayesian image analysisbayesImageS: an R package for Bayesian image analysis
bayesImageS: an R package for Bayesian image analysisMatt Moores
 
Exploratory Analysis of Multivariate Data
Exploratory Analysis of Multivariate DataExploratory Analysis of Multivariate Data
Exploratory Analysis of Multivariate DataMatt Moores
 
R package bayesImageS: Scalable Inference for Intractable Likelihoods
R package bayesImageS: Scalable Inference for Intractable LikelihoodsR package bayesImageS: Scalable Inference for Intractable Likelihoods
R package bayesImageS: Scalable Inference for Intractable LikelihoodsMatt Moores
 
Approximate Bayesian computation for the Ising/Potts model
Approximate Bayesian computation for the Ising/Potts modelApproximate Bayesian computation for the Ising/Potts model
Approximate Bayesian computation for the Ising/Potts modelMatt Moores
 
Importing satellite imagery into R from NASA and the U.S. Geological Survey
Importing satellite imagery into R from NASA and the U.S. Geological SurveyImporting satellite imagery into R from NASA and the U.S. Geological Survey
Importing satellite imagery into R from NASA and the U.S. Geological SurveyMatt Moores
 
Accelerating Pseudo-Marginal MCMC using Gaussian Processes
Accelerating Pseudo-Marginal MCMC using Gaussian ProcessesAccelerating Pseudo-Marginal MCMC using Gaussian Processes
Accelerating Pseudo-Marginal MCMC using Gaussian ProcessesMatt Moores
 
Bayesian modelling and computation for Raman spectroscopy
Bayesian modelling and computation for Raman spectroscopyBayesian modelling and computation for Raman spectroscopy
Bayesian modelling and computation for Raman spectroscopyMatt Moores
 
Final PhD Seminar
Final PhD SeminarFinal PhD Seminar
Final PhD SeminarMatt Moores
 
Precomputation for SMC-ABC with undirected graphical models
Precomputation for SMC-ABC with undirected graphical modelsPrecomputation for SMC-ABC with undirected graphical models
Precomputation for SMC-ABC with undirected graphical modelsMatt Moores
 
Pre-computation for ABC in image analysis
Pre-computation for ABC in image analysisPre-computation for ABC in image analysis
Pre-computation for ABC in image analysisMatt Moores
 
Variational Bayes
Variational BayesVariational Bayes
Variational BayesMatt Moores
 
Informative Priors for Segmentation of Medical Images
Informative Priors for Segmentation of Medical ImagesInformative Priors for Segmentation of Medical Images
Informative Priors for Segmentation of Medical ImagesMatt Moores
 

More from Matt Moores (14)

bayesImageS: an R package for Bayesian image analysis
bayesImageS: an R package for Bayesian image analysisbayesImageS: an R package for Bayesian image analysis
bayesImageS: an R package for Bayesian image analysis
 
Exploratory Analysis of Multivariate Data
Exploratory Analysis of Multivariate DataExploratory Analysis of Multivariate Data
Exploratory Analysis of Multivariate Data
 
R package bayesImageS: Scalable Inference for Intractable Likelihoods
R package bayesImageS: Scalable Inference for Intractable LikelihoodsR package bayesImageS: Scalable Inference for Intractable Likelihoods
R package bayesImageS: Scalable Inference for Intractable Likelihoods
 
Approximate Bayesian computation for the Ising/Potts model
Approximate Bayesian computation for the Ising/Potts modelApproximate Bayesian computation for the Ising/Potts model
Approximate Bayesian computation for the Ising/Potts model
 
Importing satellite imagery into R from NASA and the U.S. Geological Survey
Importing satellite imagery into R from NASA and the U.S. Geological SurveyImporting satellite imagery into R from NASA and the U.S. Geological Survey
Importing satellite imagery into R from NASA and the U.S. Geological Survey
 
Accelerating Pseudo-Marginal MCMC using Gaussian Processes
Accelerating Pseudo-Marginal MCMC using Gaussian ProcessesAccelerating Pseudo-Marginal MCMC using Gaussian Processes
Accelerating Pseudo-Marginal MCMC using Gaussian Processes
 
Bayesian modelling and computation for Raman spectroscopy
Bayesian modelling and computation for Raman spectroscopyBayesian modelling and computation for Raman spectroscopy
Bayesian modelling and computation for Raman spectroscopy
 
Final PhD Seminar
Final PhD SeminarFinal PhD Seminar
Final PhD Seminar
 
Precomputation for SMC-ABC with undirected graphical models
Precomputation for SMC-ABC with undirected graphical modelsPrecomputation for SMC-ABC with undirected graphical models
Precomputation for SMC-ABC with undirected graphical models
 
Intro to ABC
Intro to ABCIntro to ABC
Intro to ABC
 
Pre-computation for ABC in image analysis
Pre-computation for ABC in image analysisPre-computation for ABC in image analysis
Pre-computation for ABC in image analysis
 
Variational Bayes
Variational BayesVariational Bayes
Variational Bayes
 
Parallel R
Parallel RParallel R
Parallel R
 
Informative Priors for Segmentation of Medical Images
Informative Priors for Segmentation of Medical ImagesInformative Priors for Segmentation of Medical Images
Informative Priors for Segmentation of Medical Images
 

Recently uploaded

FAIR & AI Ready KGs for Explainable Predictions
FAIR & AI Ready KGs for Explainable PredictionsFAIR & AI Ready KGs for Explainable Predictions
FAIR & AI Ready KGs for Explainable PredictionsMichel Dumontier
 
The solar dynamo begins near the surface
The solar dynamo begins near the surfaceThe solar dynamo begins near the surface
The solar dynamo begins near the surfaceSérgio Sacani
 
Microbial bio Synthesis of nanoparticles.pptx
Microbial bio Synthesis of nanoparticles.pptxMicrobial bio Synthesis of nanoparticles.pptx
Microbial bio Synthesis of nanoparticles.pptxCherry
 
GBSN - Microbiology (Lab 2) Compound Microscope
GBSN - Microbiology (Lab 2) Compound MicroscopeGBSN - Microbiology (Lab 2) Compound Microscope
GBSN - Microbiology (Lab 2) Compound MicroscopeAreesha Ahmad
 
Triploidy ...............................pptx
Triploidy ...............................pptxTriploidy ...............................pptx
Triploidy ...............................pptxCherry
 
Phytogeography........................pptx
Phytogeography........................pptxPhytogeography........................pptx
Phytogeography........................pptxCherry
 
mixotrophy in cyanobacteria: a dual nutritional strategy
mixotrophy in cyanobacteria: a dual nutritional strategymixotrophy in cyanobacteria: a dual nutritional strategy
mixotrophy in cyanobacteria: a dual nutritional strategyMansiBishnoi1
 
PLANT DISEASE MANAGEMENT PRINCIPLES AND ITS IMPORTANCE
PLANT DISEASE MANAGEMENT PRINCIPLES AND ITS IMPORTANCEPLANT DISEASE MANAGEMENT PRINCIPLES AND ITS IMPORTANCE
PLANT DISEASE MANAGEMENT PRINCIPLES AND ITS IMPORTANCETALAPATI ARUNA CHENNA VYDYANAD
 
Detectability of Solar Panels as a Technosignature
Detectability of Solar Panels as a TechnosignatureDetectability of Solar Panels as a Technosignature
Detectability of Solar Panels as a TechnosignatureSérgio Sacani
 
Multi-source connectivity as the driver of solar wind variability in the heli...
Multi-source connectivity as the driver of solar wind variability in the heli...Multi-source connectivity as the driver of solar wind variability in the heli...
Multi-source connectivity as the driver of solar wind variability in the heli...Sérgio Sacani
 
Gliese 12 b: A Temperate Earth-sized Planet at 12 pc Ideal for Atmospheric Tr...
Gliese 12 b: A Temperate Earth-sized Planet at 12 pc Ideal for Atmospheric Tr...Gliese 12 b: A Temperate Earth-sized Planet at 12 pc Ideal for Atmospheric Tr...
Gliese 12 b: A Temperate Earth-sized Planet at 12 pc Ideal for Atmospheric Tr...Sérgio Sacani
 
Emergent ribozyme behaviors in oxychlorine brines indicate a unique niche for...
Emergent ribozyme behaviors in oxychlorine brines indicate a unique niche for...Emergent ribozyme behaviors in oxychlorine brines indicate a unique niche for...
Emergent ribozyme behaviors in oxychlorine brines indicate a unique niche for...Sérgio Sacani
 
National Biodiversity protection initiatives and Convention on Biological Di...
National Biodiversity protection initiatives and  Convention on Biological Di...National Biodiversity protection initiatives and  Convention on Biological Di...
National Biodiversity protection initiatives and Convention on Biological Di...PABOLU TEJASREE
 
Pests of sugarcane_Binomics_IPM_Dr.UPR.pdf
Pests of sugarcane_Binomics_IPM_Dr.UPR.pdfPests of sugarcane_Binomics_IPM_Dr.UPR.pdf
Pests of sugarcane_Binomics_IPM_Dr.UPR.pdfPirithiRaju
 
METHODS OF TRANSCRIPTOME ANALYSIS....pptx
METHODS OF TRANSCRIPTOME ANALYSIS....pptxMETHODS OF TRANSCRIPTOME ANALYSIS....pptx
METHODS OF TRANSCRIPTOME ANALYSIS....pptxCherry
 
Topography and sediments of the floor of the Bay of Bengal
Topography and sediments of the floor of the Bay of BengalTopography and sediments of the floor of the Bay of Bengal
Topography and sediments of the floor of the Bay of BengalMd Hasan Tareq
 
Structures and textures of metamorphic rocks
Structures and textures of metamorphic rocksStructures and textures of metamorphic rocks
Structures and textures of metamorphic rockskumarmathi863
 
Unveiling The Crucial Role Of Cobalt In Plant
Unveiling The Crucial Role Of Cobalt In PlantUnveiling The Crucial Role Of Cobalt In Plant
Unveiling The Crucial Role Of Cobalt In PlantHimanshu Pandey
 
Extensive Pollution of Uranus and Neptune’s Atmospheres by Upsweep of Icy Mat...
Extensive Pollution of Uranus and Neptune’s Atmospheres by Upsweep of Icy Mat...Extensive Pollution of Uranus and Neptune’s Atmospheres by Upsweep of Icy Mat...
Extensive Pollution of Uranus and Neptune’s Atmospheres by Upsweep of Icy Mat...Sérgio Sacani
 
Exomoons & Exorings with the Habitable Worlds Observatory I: On the Detection...
Exomoons & Exorings with the Habitable Worlds Observatory I: On the Detection...Exomoons & Exorings with the Habitable Worlds Observatory I: On the Detection...
Exomoons & Exorings with the Habitable Worlds Observatory I: On the Detection...Sérgio Sacani
 

Recently uploaded (20)

FAIR & AI Ready KGs for Explainable Predictions
FAIR & AI Ready KGs for Explainable PredictionsFAIR & AI Ready KGs for Explainable Predictions
FAIR & AI Ready KGs for Explainable Predictions
 
The solar dynamo begins near the surface
The solar dynamo begins near the surfaceThe solar dynamo begins near the surface
The solar dynamo begins near the surface
 
Microbial bio Synthesis of nanoparticles.pptx
Microbial bio Synthesis of nanoparticles.pptxMicrobial bio Synthesis of nanoparticles.pptx
Microbial bio Synthesis of nanoparticles.pptx
 
GBSN - Microbiology (Lab 2) Compound Microscope
GBSN - Microbiology (Lab 2) Compound MicroscopeGBSN - Microbiology (Lab 2) Compound Microscope
GBSN - Microbiology (Lab 2) Compound Microscope
 
Triploidy ...............................pptx
Triploidy ...............................pptxTriploidy ...............................pptx
Triploidy ...............................pptx
 
Phytogeography........................pptx
Phytogeography........................pptxPhytogeography........................pptx
Phytogeography........................pptx
 
mixotrophy in cyanobacteria: a dual nutritional strategy
mixotrophy in cyanobacteria: a dual nutritional strategymixotrophy in cyanobacteria: a dual nutritional strategy
mixotrophy in cyanobacteria: a dual nutritional strategy
 
PLANT DISEASE MANAGEMENT PRINCIPLES AND ITS IMPORTANCE
PLANT DISEASE MANAGEMENT PRINCIPLES AND ITS IMPORTANCEPLANT DISEASE MANAGEMENT PRINCIPLES AND ITS IMPORTANCE
PLANT DISEASE MANAGEMENT PRINCIPLES AND ITS IMPORTANCE
 
Detectability of Solar Panels as a Technosignature
Detectability of Solar Panels as a TechnosignatureDetectability of Solar Panels as a Technosignature
Detectability of Solar Panels as a Technosignature
 
Multi-source connectivity as the driver of solar wind variability in the heli...
Multi-source connectivity as the driver of solar wind variability in the heli...Multi-source connectivity as the driver of solar wind variability in the heli...
Multi-source connectivity as the driver of solar wind variability in the heli...
 
Gliese 12 b: A Temperate Earth-sized Planet at 12 pc Ideal for Atmospheric Tr...
Gliese 12 b: A Temperate Earth-sized Planet at 12 pc Ideal for Atmospheric Tr...Gliese 12 b: A Temperate Earth-sized Planet at 12 pc Ideal for Atmospheric Tr...
Gliese 12 b: A Temperate Earth-sized Planet at 12 pc Ideal for Atmospheric Tr...
 
Emergent ribozyme behaviors in oxychlorine brines indicate a unique niche for...
Emergent ribozyme behaviors in oxychlorine brines indicate a unique niche for...Emergent ribozyme behaviors in oxychlorine brines indicate a unique niche for...
Emergent ribozyme behaviors in oxychlorine brines indicate a unique niche for...
 
National Biodiversity protection initiatives and Convention on Biological Di...
National Biodiversity protection initiatives and  Convention on Biological Di...National Biodiversity protection initiatives and  Convention on Biological Di...
National Biodiversity protection initiatives and Convention on Biological Di...
 
Pests of sugarcane_Binomics_IPM_Dr.UPR.pdf
Pests of sugarcane_Binomics_IPM_Dr.UPR.pdfPests of sugarcane_Binomics_IPM_Dr.UPR.pdf
Pests of sugarcane_Binomics_IPM_Dr.UPR.pdf
 
METHODS OF TRANSCRIPTOME ANALYSIS....pptx
METHODS OF TRANSCRIPTOME ANALYSIS....pptxMETHODS OF TRANSCRIPTOME ANALYSIS....pptx
METHODS OF TRANSCRIPTOME ANALYSIS....pptx
 
Topography and sediments of the floor of the Bay of Bengal
Topography and sediments of the floor of the Bay of BengalTopography and sediments of the floor of the Bay of Bengal
Topography and sediments of the floor of the Bay of Bengal
 
Structures and textures of metamorphic rocks
Structures and textures of metamorphic rocksStructures and textures of metamorphic rocks
Structures and textures of metamorphic rocks
 
Unveiling The Crucial Role Of Cobalt In Plant
Unveiling The Crucial Role Of Cobalt In PlantUnveiling The Crucial Role Of Cobalt In Plant
Unveiling The Crucial Role Of Cobalt In Plant
 
Extensive Pollution of Uranus and Neptune’s Atmospheres by Upsweep of Icy Mat...
Extensive Pollution of Uranus and Neptune’s Atmospheres by Upsweep of Icy Mat...Extensive Pollution of Uranus and Neptune’s Atmospheres by Upsweep of Icy Mat...
Extensive Pollution of Uranus and Neptune’s Atmospheres by Upsweep of Icy Mat...
 
Exomoons & Exorings with the Habitable Worlds Observatory I: On the Detection...
Exomoons & Exorings with the Habitable Worlds Observatory I: On the Detection...Exomoons & Exorings with the Habitable Worlds Observatory I: On the Detection...
Exomoons & Exorings with the Habitable Worlds Observatory I: On the Detection...
 

Bayesian Inference and Uncertainty Quantification for Inverse Problems

  • 1. Bayesian Inference and Uncertainty Quantification for Inverse Problems Matt Moores @MooresMt https://uow.edu.au/~mmoores/ Centre for Environmental Informatics (CEI), University of Wollongong, NSW, Australia TIDE Hub Seminar Series joint with Matthew Berry, Mark Nelson, Brian Monaghan and Raymond Longbottom 1 / 19
  • 2. Introduction Physical model F t; ~ θ Unknown parameters ~ θ = {θ1, . . . , θk} Initial conditions F0 and derivatives ∂F ∂t , etc. 2 / 19
  • 3. Introduction Physical model F t; ~ θ Unknown parameters ~ θ = {θ1, . . . , θk} Initial conditions F0 and derivatives ∂F ∂t , etc. Observed data yt at times t = 1, . . . , T yt = F t; ~ θ + εt where εt is random noise. 2 / 19
  • 5. Example Growth curve: dF dt = −βγt log γ Solution: F ti; ~ θ = α − βγti at time points t1, . . . , tn with initial condition F0 = α − β. 3 / 19
  • 6. Example Growth curve: dF dt = −βγt log γ Solution: F ti; ~ θ = α − βγti at time points t1, . . . , tn with initial condition F0 = α − β. Noisy observations: yi = F(ti; α, β, γ) + εi where εi ∼ N(0, σ2 ) is additive Gaussian noise. Unknown parameters ~ θ = {α, β, γ, σ2 } 3 / 19
  • 7. Observed Data for n = 27 dugongs 4 / 19
  • 8. Nonlinear Least Squares nlm - nls(y ~ alpha - beta * gammˆt, data=dat, start=list(alpha=1, beta=1, gamm=0.9)) summary(nlm) ## ## Formula: y ~ alpha - beta * gamm^t ## ## Parameters: ## Estimate Std. Error t value Pr(|t|) ## alpha 2.65807 0.06151 43.21 2e-16 *** ## beta 0.96352 0.06968 13.83 6.3e-13 *** ## gamm 0.87146 0.02460 35.42 2e-16 *** ## --- ## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 ## ## Residual standard error: 0.09525 on 24 degrees of freedom 5 / 19
  • 9. Drawbacks No guarantee of convexity: will only find the nearest local optimum results are completely dependent on initialization 6 / 19
  • 10. Drawbacks No guarantee of convexity: will only find the nearest local optimum results are completely dependent on initialization Standard errors are underestimated: 95% profile CI for α and β only achieve ~83% coverage 6 / 19
  • 11. Inverse Problem Find parameters ~ θ consistent with ~ y 7 / 19
  • 12. Inverse Problem Find parameters ~ θ consistent with ~ y Physical model F : Θ → Y doesn’t need to be invertible: although lack of identifiability for ~ θ can create problems we don’t need F in closed form — can use numeric solver for DE 7 / 19
  • 13. Inverse Problem Find parameters ~ θ consistent with ~ y Physical model F : Θ → Y doesn’t need to be invertible: although lack of identifiability for ~ θ can create problems we don’t need F in closed form — can use numeric solver for DE Likelihood L(~ y | ~ θ) is based on E[Yi | ~ θ] = F ti; ~ θ as well as distribution of random noise (not necessarily Gaussian) 7 / 19
  • 14. Bayesian Inference We assign prior distributions to the parameters ~ θ: π(log{α}) ∼ N(0, 100) π(log{β}) ∼ N(0, 100) π(logit{γ}) ∼ N(0, 100) π 1 σ2 ∼ Ga ν0 2 , 2 ν0s2 0 ! 8 / 19 Astfalck, Cripps, Gosling, Hodkiewicz Milne (2018) Ocean Engineering 161: 268–276.
  • 15. Bayesian Inference We assign prior distributions to the parameters ~ θ: π(log{α}) ∼ N(0, 100) π(log{β}) ∼ N(0, 100) π(logit{γ}) ∼ N(0, 100) π 1 σ2 ∼ Ga ν0 2 , 2 ν0s2 0 ! Then the joint posterior distribution is: π ~ θ | ~ y = L(~ y | ~ θ)π(~ θ) π(~ y) where the normalising constant π(~ y) = R ∞ 0 R ∞ 0 R ∞ 0 R 1 0 L(~ y | α, β, γ, σ) dπ(γ)dπ(β)dπ(α)dπ(σ) is intractable. 8 / 19 Astfalck, Cripps, Gosling, Hodkiewicz Milne (2018) Ocean Engineering 161: 268–276.
  • 16. Markov Chain Monte Carlo Algorithm 1 Random-walk Metropolis sampler for π ~ θ | ~ y 1: Initialize α0 ∼ π(log{α}), β0 ∼ π(log{β}), γ0 ∼ π(logit{γ}), σ0 ∼ π 1 σ2 2: Solve ~ µ0 = ~ F ~ t; α0, β0, γ0 and calculate `0 = log L(~ y | ~ µ0, σ2 0) 3: for iterations j = 1, . . . , J do 4: Propose log{α∗} ∼ N(log{αj−1}, σ2 α), log{β∗} ∼ N(log{βj−1}, σ2 β), logit{γ∗} ∼ N(logit{γj−1}, σ2 γ) 5: Solve ~ µ∗ = ~ F ~ t; α∗, β∗, γ∗ 6: Propose σ∗ ∼ π 1 σ2 | ~ µ∗,~ y and calculate `∗ = log L(~ y | ~ µ∗, σ2 ∗) 7: Calculate ρt = exp{`∗}π(~ θ∗)q(~ θj−1 | ~ θ∗) exp{`j−1}π(~ θj−1)q(~ θ∗ | ~ θj−1) 8: Accept ~ θj = ~ θ∗ with probability min{ρt, 1} else ~ θj = ~ θj−1 9: end for 9 / 19
  • 20. Thermogravimetric Analysis The TGA model for a single reaction involves the Arrhenius equations: dM dt = −MA exp − E RT dT dt = α where M is mass fraction, T is temperature, R is the ideal gas constant. The initial mass M0, the initial temperature T0, and the heating rate α are experimentally controlled. The unknown parameters ~ θ are A pre-exponential factor E activation energy σ2 variance of the additive Gaussian noise 13 / 19
  • 21. Reparameterisation We have good prior information for E (physically interpretable) and σ2 (measurement noise), but not for A. The distributions for A and E are also very highly correlated, slowing the mixing of MCMC. 14 / 19
  • 22. Reparameterisation We have good prior information for E (physically interpretable) and σ2 (measurement noise), but not for A. The distributions for A and E are also very highly correlated, slowing the mixing of MCMC. Instead, we reparameterise the model in terms of the temperature, Tm, at which the rate dM dt is maximised: A exp − E RTm = E RT2 m α In our MCMC algorithm, we first propose q(E∗ | Ej−1), then propose q(Tm∗ | Tm,j−1). The value of A∗ can then be obtained from the equation above. We solve ~ F ~ t; E∗, A∗ numerically using a Runge-Kutta method. 14 / 19
  • 23. Posterior Distributions 634.6 634.7 634.8 634.9 635.0 635.1 635.2 635.3 Tm 0 1 2 3 4 5 84750 85000 85250 85500 85750 86000 86250 86500 86750 E 0.0000 0.0002 0.0004 0.0006 0.0008 0.0010 0.0012 0.0014 0.0016 0.08 0.10 0.12 0.14 0 10 20 30 40 50 2400000 2600000 2800000 3000000 3200000 3400000 3600000 A 0.0000000 0.0000005 0.0000010 0.0000015 0.0000020 0.0000025 15 / 19
  • 24. Posterior for Functions of ~ θ A function of a random variable is also a random variable. Now that we have random samples from our posterior π(A, E, Tm, σ2 | ~ y), we can use our model to obtain predictions for any measurable function of the parameters, g(A, E, Tm). 16 / 19
  • 25. Posterior for Functions of ~ θ A function of a random variable is also a random variable. Now that we have random samples from our posterior π(A, E, Tm, σ2 | ~ y), we can use our model to obtain predictions for any measurable function of the parameters, g(A, E, Tm). For example, the critical length of a stockpile: Lcr = g(A, E, Tm) = K v u u texp n E RT2 m o A RT2 m E E[g(A, E, Tm) | ~ y] = Z 1000 0 Z ∞ 0 Z ∞ 0 g(A, E, Tm) dπ(A, E, Tm | ~ y) ≈ J X j=1 g(Aj, Ej, Tj) where K is a constant and {Aj, Ej, Tj}J j=1 are the MCMC samples (after discarding burn-in). 16 / 19
  • 27. Advanced Methods Parallel and distributed computation Sequential updating of π(~ θ | ~ y) and streaming inference Spatio-temporal modelling of F(x, y, z, t; ~ θ) Emulation of the forward map Θ → Y (e.g. using artificial neural networks) Approximating the likelihood L (e.g. ABC) Accounting for multi-modality (e.g. ill-posed problems) Accounting for error in the numerical solver (probabilistic numerics) Accounting for model misspecification (discrepancy function) Multi-level Monte Carlo methods Hamiltonian Monte Carlo 18 / 19
  • 28. Further Reading R. J. Longbottom, B. J. Monaghan, D. J. Pinson, N. A. S. Webster S. J. Chew In situ Phase Analysis during Self-sintering of BOS Filter Cake for Improved Recycling. ISIJ International 60(11): 2436–2445, 2020. A. M. Stuart Inverse problems: A Bayesian perspective. Acta Numerica 19: 451—559, 2010. L.C. Astfalck, E.J. Cripps, J.P. Gosling, M.R. Hodkiewicz I.A. Milne Expert elicitation of directional metocean parameters. Ocean Engineering 161: 268–276, 2018. B. P. Carlin A. E. Gelfand An iterative Monte Carlo method for nonconjugate Bayesian analysis. Statistics Computing 1(2): 119–128, 1991. 19 / 19