SlideShare a Scribd company logo
1 of 29
Download to read offline
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
an introduction to
Approximate Bayesian Computation
Matt Moores
Research Fellow
Department of Statistics
University of Warwick
Warwick ML Club
June 12, 2017
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
Motivation
Inference for a parameter θ when it is:
impossible
or very expensive
to evaluate the likelihood p(y|θ)
ABC is a likelihood-free method for approximating
the posterior distribution
π(θ|y)
by generating pseudo-data from the model:
w ∼ f(·|θ)
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
Likelihood-free rejection sampler
Algorithm 1 Likelihood-free rejection sampler
1: Draw parameter value θ ∼ π(θ)
2: Generate w ∼ f(·|θ )
3: if w = y (the observed data) then
4: accept θ
5: end if
But if the observations y are continuous
(or the space y ∈ Y is enormous)
then P(w = y) ≈ 0
Tavar´e, Balding, Griffith & Donnelly (1997) Genetics 145(2)
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
ABC tolerance
accept θ if δ(w, y) <
where
> 0 is the tolerance level
δ(·, ·) is a distance function
(for an appropriate choice of norm)
Inference is more exact when is close to zero. but
more proposed θ are rejected
(tradeoff between accuracy & computational cost)
Pritchard, Seielstad, Perez-Lezaun & Feldman (1999) Mol. Biol. Evol. 16(12)
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
Summary statistics
Computing δ(w, y) for w1, . . . , wn and y1, . . . , yn
can be very expensive for large n
Instead, compute summary statistics s(y)
e.g. sufficient statistics
(only available for exponential family)
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
Sufficient statistics
Fisher-Neyman factorisation theorem:
if s(y) is sufficient for θ
then p(y|θ) = f(y) g (s(y)|θ)
only applies to Potts, Ising, exponential random
graph models (ERGM)
otherwise, selection of suitable summary
statistics can be a very difficult problem
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
ABC rejection sampler
Algorithm 2 ABC rejection sampler
1: for all iterations t ∈ 1 . . . T do
2: Draw independent proposal θ ∼ π(θ)
3: Generate w ∼ f(·|θ )
4: if s(w) − s(y) < then
5: set θt ← θ
6: else
7: set θt ← θt−1
8: end if
9: end for
Approximates π(θ|y) by π (θ | s(w) − s(y) < )
Marin, Pudlo, Robert & Ryder (2012) Stat. Comput. 22(6)
Marin & Robert (2014) Bayesian Essentials with R §8.3
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
A trivial example
Gaussian with unknown variance:
y ∼ N(1, σ2
)
normalising constant:
Z(σ2
) = (2πσ2
)−n
2
natural conjugate prior:
π 1
σ2 ∼ Ga ν0
2
,
ν0ψ2
0
2
sufficient statistic:
s(y) = 1
n
n
i=1 (yi − 1)2
posterior is analytically tractable:
π 1
σ2 | y ∼ Ga ν0+n
2
,
ν0ψ2
0+ns(y)
2
∴ no need for ABC (nor MCMC) in practice
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
R code
π(τ|y)
Density
0 1 2 3 4 5 6 7
0.00.10.20.30.4
§
y ← rnorm (n=5, mean=1, sd=2/3)
n ← length ( y )
s sq ← sum(( y −1)ˆ2)/n
post nu ← nu0 + n
post ssd ← ( nu0 ∗ s0 ˆ2 + n∗ s sq )/2
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
now with ABC
π(τ)
0 1 2 3 4 5 6 7
0.00.51.01.5
πε(τ | δ(s(w), s(y)) < ε)
0 1 2 3 4 5 6 7
0.00.10.20.30.4
§
prop tau ← rgamma(10000 , nu0/2 , 0.5 ∗nu0∗ s0 ˆ2)
pseudo ← rnorm (n∗ 10000 , 1 , 1/ sqrt ( prop tau ))
pseudoMx ← matrix ( pseudo , nrow=10000, ncol=n)
pseudoSSD ← rowSums (( pseudoMx − 1)ˆ2)/n
ps norm ← abs ( pseudoSSD − s sq )
e p s i l o n ← sort ( ps norm ) [ 2 0 0 0 ]
prop keep ← prop tau [ ps norm <= e p s i l o n ]
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
choice of
0 1 2 3 4 5 6 7
0.00.20.40.6
(a) = 8.430
0 1 2 3 4 5 6 7
0.00.10.20.30.4
(b) = 1.427
0 1 2 3 4 5 6 7
0.00.20.40.6
(c) = 0.011
0 1 2 3 4 5 6 7
0.00.40.8
(d) = 0.001
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
concentration of measure
0 1 2 3 4
0.00.40.8
(a) n = 25, = 0.149
0 1 2 3 4
0.01.02.03.0
(b) n = 500, = 0.025
0 1 2 3 4
02468
(c) n = 104
, = 0.011
0 1 2 3 4
020406080100
(d) n = 106
, = 0.001
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
Metropolis-Hastings proposals
Algorithm 3 ABC-MCMC
1: Initialise θ0 ∼ π(θ)
2: for all iterations t ∈ 1 . . . T do
3: Draw proposal θ ∼ q(· | θt−1)
4: Generate w ∼ f(·|θ )
5: Draw u ∼ Unif(0, 1)
6: if u < π(θ )q(θt−1|θ )
π(θt−1)q(θ |θt−1) and s(w) − s(y) < then
7: set θt ← θ
8: else
9: set θt ← θt−1
10: end if
11: end for
Unfortunately, this algorithm is prone to getting ”stuck”
Marjoram, Molitor, Plagnol & Tavar´e (2003) PNAS 100(26)
Lee & Latuszy´nski (2014) Biometrika 101(3)
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
Sequential Monte Carlo
Algorithm 4 ABC-SMC
1: Draw N particles θi ∼ π0(θ)
2: Draw N × M sets of pseudo-data wi,m ∼ f(·|θi)
3: repeat
4: Adaptively select ABC tolerance t
5: Update importance weights λi for each particle
6: if effective sample size (ESS) < Nmin then
7: Resample particles according to their weights
8: end if
9: Update particles using Metropolis-Hastings step
(with adaptive proposal bandwidth σ2
t )
10: until
naccept
N < 0.015 or t < 10−9 or t ≥ 100
Targets a sequence of distributions π t (θ | s(w) − s(y) < t)
such that 1 > 2 > · · · > T
Drovandi & Pettitt (2011) Biometrics 67(1)
Del Moral, Doucet & Jasra (2012) Stat. Comput. 22(5)
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
hidden Markov random field
Joint distribution of observed pixel intensities
y = (y1, . . . , yn) ∈ Rn and latent labels
z = (z1, . . . , zn) ∈ {1, . . . , k}n:
Pr(y, z | µ, σ2
, β) = L(y|µ, σ2
, z)π(z|β) (1)
Additive Gaussian noise:
yi | zi =j
iid
∼ N µj, σ2
j (2)
Potts model:
π(zi | zi, β) =
exp {β i∼ δ(zi, z )}
k
j=1 exp {β i∼ δ(j, z )}
(3)
Potts (1952) Proceedings of the Cambridge Philosophical Society 48(1)
Winkler (2003) Image Analysis, Random Fields and MCMC Methods, 2nd
ed.
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
Inverse Temperature
(e) β = 0.1 (f) β = 0.5 (g) β = 0.85
(h) β = 0.95 (i) β = 1.005 (j) β = 1.15
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
Doubly-intractable posterior
p(β|z) =
C−1(β)eβS(z)π(β)
β C−1(β)eβS(z)π(dβ)
(4)
The normalising constant has computational complexity O(nkn),
since it involves a sum over all possible combinations of the labels
z ∈ Z:
C(β) =
z∈Z
eβS(z)
(5)
S(z) is the sufficient statistic of the Potts model:
S(z) =
i∼ ∈E
δ(zi, z ) (6)
where E is the set of all unique neighbour pairs.
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
bayesImageS
An R package for Bayesian image segmentation
using the hidden Potts model:
RcppArmadillo for fast computation in C++
OpenMP for parallelism
§
l i b r a r y ( bayesImageS )
p r i o r s ← l i s t ("k"=3,"mu"=rep (0 ,3) , "mu.sd"=sigma ,
"sigma"=sigma , "sigma.nu"=c (1 ,1 ,1) , "beta"=c (0 ,3))
mh ← l i s t ( algorithm="pseudo" , bandwidth =0.2)
r e s u l t ← mcmcPotts ( y , neigh , block ,NULL,
55000 ,5000 , p r i o r s ,mh)
Eddelbuettel & Sanderson (2014) RcppArmadillo: Accelerating R with
high-performance C++ linear algebra. CSDA 71
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
Bayesian computational methods
bayesImageS supports methods for updating the latent labels z:
Chequerboard updating (Winkler 2003)
Swendsen-Wang (1987)
and also methods for updating the inverse temperature β:
Pseudolikelihood (Ryd´en & Titterington 1998)
Path Sampling (Gelman & Meng 1998)
Exchange Algorithm (Murray, Ghahramani & MacKay 2006)
Approximate Bayesian Computation (Grelaud et al. 2009)
Sequential Monte Carlo (SMC-ABC)
(Del Moral, Doucet & Jasra 2012)
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
Lake Menteith, Scotland
0.0 0.2 0.4 0.6 0.8 1.0
0.00.20.40.60.81.0
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
ABC-SMC for the hidden Potts model
§
l i b r a r y ( bayess )
data ( Menteith )
y ← as . matrix ( Menteith )
l i b r a r y ( bayesImageS )
mask ← matrix (1 , nrow=nrow( y ) , ncol=ncol ( y ))
neigh ← getNeighbors (mask , c (2 ,2 ,0 ,0))
block ← getBlocks (mask , 2)
p r i o r s ← l i s t ( k=6, mu=rep (256/2 , 6) ,
mu. sd=rep (256/ 6 ,6) , sigma=rep (256/ 6 ,6) ,
sigma . nu=rep (6 , 6) , beta ← c (0 ,2))
r e s ← smcPotts ( as . vector ( y ) , neigh , block ,
param=l i s t ( npart =2000, nstat =5) , p r i o r s=p r i o r s )
6h 15min for 100 SMC iterations (N=2000, M=5)
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
ABC tolerance
SMC iteration
εt
0 20 40 60 80 100
02000400060008000100001200014000
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
approximate posterior ( t = 37.5)
β
Density
1.280 1.285 1.290 1.295 1.300 1.305 1.310 1.315
0204060
Equivalent to 120,801 iterations of the ABC rejection sampler
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
Effective Sample Size
SMC iteration
ESS
0 20 40 60 80 100
800100012001400160018002000
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
image segmentation
(a) Original image (b) Potts labels
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
Summary
ABC is a method for likelihood-free inference
It enables inference for models that are
otherwise computationally intractable
Main components of ABC:
π(θ) proposal density for θ
f(·|θ) generative model for w
tolerance level
δ(·, ·) distance function
s(y) summary statistics
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
References I
M. Moores, A. N. Pettitt & K. Mengersen
Scalable Bayesian inference for the inverse temperature of a hidden Potts
model.
arXiv:1503.08066 [stat.CO], 2015.
M. Moores, C. C. Drovandi, K. Mengersen & C. P. Robert
Pre-processing for approximate Bayesian computation in image analysis.
Statistics & Computing 25(1): 23–33, 2015.
J.-M. Marin, P. Pudlo, C. P. Robert & R. Ryder
Approximate Bayesian computational methods.
Statistics & Computing, 22(6): 1167–80, 2012.
A. Grelaud, C. P. Robert, J.-M. Marin, F. Rodolphe & J.-F. Taly
ABC likelihood-free methods for model choice in Gibbs random fields.
Bayesian Analysis, 4(2): 317–36, 2009.
J.-M. Marin & C. P. Robert
Bayesian Essentials with R
Springer-Verlag, 2014.
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
References II
A. Lee & K. Latuszy´nski
Variance bounding and geometric ergodicity of Markov chain Monte Carlo
kernels for approximate Bayesian computation
Biometrika 101(3): 655–671, 2014.
P. Del Moral, A. Doucet & A. Jasra
An adaptive sequential Monte Carlo method for approximate Bayesian
computation.
Statistics & Computing, 22(5): 1009–20, 2012.
C. C. Drovandi & A. N. Pettitt
Estimation of Parameters for Macroparasite Population Evolution Using
Approximate Bayesian Computation
Biometrics 67(1): 225–233, 2011.
P. Marjoram, J. Molitor, V. Plagnol & S. Tavar´e
Markov chain Monte Carlo without likelihoods.
Proc. Natl Acad. Sci. USA, 100(26): 15324–15328, 2003.
Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion
References III
J. Pritchard, M. Seielstad, A. Perez-Lezaun & M. Feldman
Population Growth of Human Y Chromosomes: A Study of Y
Chromosome Microsatellites.
Mol. Biol. Evol. 16(12): 1791–98, 1999.
S. Tavar´e, D. Balding, R, Griffiths & P. Donnelly
Inferring coalescence times from DNA sequence data.
Genetics, 145(2): 505–18, 1997.
R. B. Potts
Some generalized order-disorder transformations.
Proc. Cambridge Philosophical Society, 48(1): 106–109, 1952.
G. Winkler
Image analysis, random fields and Markov chain Monte Carlo methods
2nd
ed., Springer-Verlag, 2003.
D. Eddelbuettel
Seamless R and C++ integration with Rcpp
Springer-Verlag, 2013.

More Related Content

What's hot

Poster for Bayesian Statistics in the Big Data Era conference
Poster for Bayesian Statistics in the Big Data Era conferencePoster for Bayesian Statistics in the Big Data Era conference
Poster for Bayesian Statistics in the Big Data Era conferenceChristian Robert
 
Inference in generative models using the Wasserstein distance [[INI]
Inference in generative models using the Wasserstein distance [[INI]Inference in generative models using the Wasserstein distance [[INI]
Inference in generative models using the Wasserstein distance [[INI]Christian Robert
 
Intro to Approximate Bayesian Computation (ABC)
Intro to Approximate Bayesian Computation (ABC)Intro to Approximate Bayesian Computation (ABC)
Intro to Approximate Bayesian Computation (ABC)Umberto Picchini
 
NCE, GANs & VAEs (and maybe BAC)
NCE, GANs & VAEs (and maybe BAC)NCE, GANs & VAEs (and maybe BAC)
NCE, GANs & VAEs (and maybe BAC)Christian Robert
 
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
 
Monte Carlo in Montréal 2017
Monte Carlo in Montréal 2017Monte Carlo in Montréal 2017
Monte Carlo in Montréal 2017Christian Robert
 
Maximum likelihood estimation of regularisation parameters in inverse problem...
Maximum likelihood estimation of regularisation parameters in inverse problem...Maximum likelihood estimation of regularisation parameters in inverse problem...
Maximum likelihood estimation of regularisation parameters in inverse problem...Valentin De Bortoli
 
ABC with data cloning for MLE in state space models
ABC with data cloning for MLE in state space modelsABC with data cloning for MLE in state space models
ABC with data cloning for MLE in state space modelsUmberto Picchini
 
accurate ABC Oliver Ratmann
accurate ABC Oliver Ratmannaccurate ABC Oliver Ratmann
accurate ABC Oliver Ratmannolli0601
 
ABC convergence under well- and mis-specified models
ABC convergence under well- and mis-specified modelsABC convergence under well- and mis-specified models
ABC convergence under well- and mis-specified modelsChristian Robert
 
ABC based on Wasserstein distances
ABC based on Wasserstein distancesABC based on Wasserstein distances
ABC based on Wasserstein distancesChristian Robert
 
Patch Matching with Polynomial Exponential Families and Projective Divergences
Patch Matching with Polynomial Exponential Families and Projective DivergencesPatch Matching with Polynomial Exponential Families and Projective Divergences
Patch Matching with Polynomial Exponential Families and Projective DivergencesFrank Nielsen
 
Coordinate sampler : A non-reversible Gibbs-like sampler
Coordinate sampler : A non-reversible Gibbs-like samplerCoordinate sampler : A non-reversible Gibbs-like sampler
Coordinate sampler : A non-reversible Gibbs-like samplerChristian Robert
 
H2O World - Consensus Optimization and Machine Learning - Stephen Boyd
H2O World - Consensus Optimization and Machine Learning - Stephen BoydH2O World - Consensus Optimization and Machine Learning - Stephen Boyd
H2O World - Consensus Optimization and Machine Learning - Stephen BoydSri Ambati
 
Inference for stochastic differential equations via approximate Bayesian comp...
Inference for stochastic differential equations via approximate Bayesian comp...Inference for stochastic differential equations via approximate Bayesian comp...
Inference for stochastic differential equations via approximate Bayesian comp...Umberto Picchini
 
Accelerated approximate Bayesian computation with applications to protein fol...
Accelerated approximate Bayesian computation with applications to protein fol...Accelerated approximate Bayesian computation with applications to protein fol...
Accelerated approximate Bayesian computation with applications to protein fol...Umberto Picchini
 
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
 

What's hot (20)

Poster for Bayesian Statistics in the Big Data Era conference
Poster for Bayesian Statistics in the Big Data Era conferencePoster for Bayesian Statistics in the Big Data Era conference
Poster for Bayesian Statistics in the Big Data Era conference
 
Inference in generative models using the Wasserstein distance [[INI]
Inference in generative models using the Wasserstein distance [[INI]Inference in generative models using the Wasserstein distance [[INI]
Inference in generative models using the Wasserstein distance [[INI]
 
ABC-Gibbs
ABC-GibbsABC-Gibbs
ABC-Gibbs
 
Intro to Approximate Bayesian Computation (ABC)
Intro to Approximate Bayesian Computation (ABC)Intro to Approximate Bayesian Computation (ABC)
Intro to Approximate Bayesian Computation (ABC)
 
NCE, GANs & VAEs (and maybe BAC)
NCE, GANs & VAEs (and maybe BAC)NCE, GANs & VAEs (and maybe BAC)
NCE, GANs & VAEs (and maybe BAC)
 
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
 
Monte Carlo in Montréal 2017
Monte Carlo in Montréal 2017Monte Carlo in Montréal 2017
Monte Carlo in Montréal 2017
 
Maximum likelihood estimation of regularisation parameters in inverse problem...
Maximum likelihood estimation of regularisation parameters in inverse problem...Maximum likelihood estimation of regularisation parameters in inverse problem...
Maximum likelihood estimation of regularisation parameters in inverse problem...
 
ABC with data cloning for MLE in state space models
ABC with data cloning for MLE in state space modelsABC with data cloning for MLE in state space models
ABC with data cloning for MLE in state space models
 
accurate ABC Oliver Ratmann
accurate ABC Oliver Ratmannaccurate ABC Oliver Ratmann
accurate ABC Oliver Ratmann
 
ABC convergence under well- and mis-specified models
ABC convergence under well- and mis-specified modelsABC convergence under well- and mis-specified models
ABC convergence under well- and mis-specified models
 
ABC based on Wasserstein distances
ABC based on Wasserstein distancesABC based on Wasserstein distances
ABC based on Wasserstein distances
 
Patch Matching with Polynomial Exponential Families and Projective Divergences
Patch Matching with Polynomial Exponential Families and Projective DivergencesPatch Matching with Polynomial Exponential Families and Projective Divergences
Patch Matching with Polynomial Exponential Families and Projective Divergences
 
Coordinate sampler : A non-reversible Gibbs-like sampler
Coordinate sampler : A non-reversible Gibbs-like samplerCoordinate sampler : A non-reversible Gibbs-like sampler
Coordinate sampler : A non-reversible Gibbs-like sampler
 
H2O World - Consensus Optimization and Machine Learning - Stephen Boyd
H2O World - Consensus Optimization and Machine Learning - Stephen BoydH2O World - Consensus Optimization and Machine Learning - Stephen Boyd
H2O World - Consensus Optimization and Machine Learning - Stephen Boyd
 
Inference for stochastic differential equations via approximate Bayesian comp...
Inference for stochastic differential equations via approximate Bayesian comp...Inference for stochastic differential equations via approximate Bayesian comp...
Inference for stochastic differential equations via approximate Bayesian comp...
 
talk MCMC & SMC 2004
talk MCMC & SMC 2004talk MCMC & SMC 2004
talk MCMC & SMC 2004
 
Accelerated approximate Bayesian computation with applications to protein fol...
Accelerated approximate Bayesian computation with applications to protein fol...Accelerated approximate Bayesian computation with applications to protein fol...
Accelerated approximate Bayesian computation with applications to protein fol...
 
QMC Program: Trends and Advances in Monte Carlo Sampling Algorithms Workshop,...
QMC Program: Trends and Advances in Monte Carlo Sampling Algorithms Workshop,...QMC Program: Trends and Advances in Monte Carlo Sampling Algorithms Workshop,...
QMC Program: Trends and Advances in Monte Carlo Sampling Algorithms Workshop,...
 
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
 

Similar to Approximate Bayesian computation for the Ising/Potts model

Stratified sampling and resampling for approximate Bayesian computation
Stratified sampling and resampling for approximate Bayesian computationStratified sampling and resampling for approximate Bayesian computation
Stratified sampling and resampling for approximate Bayesian computationUmberto Picchini
 
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
 
Workshop in honour of Don Poskitt and Gael Martin
Workshop in honour of Don Poskitt and Gael MartinWorkshop in honour of Don Poskitt and Gael Martin
Workshop in honour of Don Poskitt and Gael MartinChristian Robert
 
Stratified Monte Carlo and bootstrapping for approximate Bayesian computation
Stratified Monte Carlo and bootstrapping for approximate Bayesian computationStratified Monte Carlo and bootstrapping for approximate Bayesian computation
Stratified Monte Carlo and bootstrapping for approximate Bayesian computationUmberto Picchini
 
ABC short course: model choice chapter
ABC short course: model choice chapterABC short course: model choice chapter
ABC short course: model choice chapterChristian Robert
 
Tensor Completion for PDEs with uncertain coefficients and Bayesian Update te...
Tensor Completion for PDEs with uncertain coefficients and Bayesian Update te...Tensor Completion for PDEs with uncertain coefficients and Bayesian Update te...
Tensor Completion for PDEs with uncertain coefficients and Bayesian Update te...Alexander Litvinenko
 
random forests for ABC model choice and parameter estimation
random forests for ABC model choice and parameter estimationrandom forests for ABC model choice and parameter estimation
random forests for ABC model choice and parameter estimationChristian Robert
 
Asymptotics of ABC, lecture, Collège de France
Asymptotics of ABC, lecture, Collège de FranceAsymptotics of ABC, lecture, Collège de France
Asymptotics of ABC, lecture, Collège de FranceChristian Robert
 
Columbia workshop [ABC model choice]
Columbia workshop [ABC model choice]Columbia workshop [ABC model choice]
Columbia workshop [ABC model choice]Christian Robert
 
Considerate Approaches to ABC Model Selection
Considerate Approaches to ABC Model SelectionConsiderate Approaches to ABC Model Selection
Considerate Approaches to ABC Model SelectionMichael Stumpf
 
Hands-On Algorithms for Predictive Modeling
Hands-On Algorithms for Predictive ModelingHands-On Algorithms for Predictive Modeling
Hands-On Algorithms for Predictive ModelingArthur Charpentier
 
ABC short course: survey chapter
ABC short course: survey chapterABC short course: survey chapter
ABC short course: survey chapterChristian Robert
 

Similar to Approximate Bayesian computation for the Ising/Potts model (20)

Stratified sampling and resampling for approximate Bayesian computation
Stratified sampling and resampling for approximate Bayesian computationStratified sampling and resampling for approximate Bayesian computation
Stratified sampling and resampling for approximate Bayesian computation
 
ABC-Gibbs
ABC-GibbsABC-Gibbs
ABC-Gibbs
 
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
 
ABC workshop: 17w5025
ABC workshop: 17w5025ABC workshop: 17w5025
ABC workshop: 17w5025
 
Workshop in honour of Don Poskitt and Gael Martin
Workshop in honour of Don Poskitt and Gael MartinWorkshop in honour of Don Poskitt and Gael Martin
Workshop in honour of Don Poskitt and Gael Martin
 
Stratified Monte Carlo and bootstrapping for approximate Bayesian computation
Stratified Monte Carlo and bootstrapping for approximate Bayesian computationStratified Monte Carlo and bootstrapping for approximate Bayesian computation
Stratified Monte Carlo and bootstrapping for approximate Bayesian computation
 
Automatic bayesian cubature
Automatic bayesian cubatureAutomatic bayesian cubature
Automatic bayesian cubature
 
ABC-Gibbs
ABC-GibbsABC-Gibbs
ABC-Gibbs
 
ABC short course: model choice chapter
ABC short course: model choice chapterABC short course: model choice chapter
ABC short course: model choice chapter
 
Multivariate Methods Assignment Help
Multivariate Methods Assignment HelpMultivariate Methods Assignment Help
Multivariate Methods Assignment Help
 
Tensor Completion for PDEs with uncertain coefficients and Bayesian Update te...
Tensor Completion for PDEs with uncertain coefficients and Bayesian Update te...Tensor Completion for PDEs with uncertain coefficients and Bayesian Update te...
Tensor Completion for PDEs with uncertain coefficients and Bayesian Update te...
 
random forests for ABC model choice and parameter estimation
random forests for ABC model choice and parameter estimationrandom forests for ABC model choice and parameter estimation
random forests for ABC model choice and parameter estimation
 
asymptotics of ABC
asymptotics of ABCasymptotics of ABC
asymptotics of ABC
 
4th Semester CS / IS (2013-June) Question Papers
4th Semester CS / IS (2013-June) Question Papers 4th Semester CS / IS (2013-June) Question Papers
4th Semester CS / IS (2013-June) Question Papers
 
Edinburgh, Bayes-250
Edinburgh, Bayes-250Edinburgh, Bayes-250
Edinburgh, Bayes-250
 
Asymptotics of ABC, lecture, Collège de France
Asymptotics of ABC, lecture, Collège de FranceAsymptotics of ABC, lecture, Collège de France
Asymptotics of ABC, lecture, Collège de France
 
Columbia workshop [ABC model choice]
Columbia workshop [ABC model choice]Columbia workshop [ABC model choice]
Columbia workshop [ABC model choice]
 
Considerate Approaches to ABC Model Selection
Considerate Approaches to ABC Model SelectionConsiderate Approaches to ABC Model Selection
Considerate Approaches to ABC Model Selection
 
Hands-On Algorithms for Predictive Modeling
Hands-On Algorithms for Predictive ModelingHands-On Algorithms for Predictive Modeling
Hands-On Algorithms for Predictive Modeling
 
ABC short course: survey chapter
ABC short course: survey chapterABC short course: survey chapter
ABC short course: survey chapter
 

More from Matt Moores

Bayesian Inference and Uncertainty Quantification for Inverse Problems
Bayesian Inference and Uncertainty Quantification for Inverse ProblemsBayesian Inference and Uncertainty Quantification for Inverse Problems
Bayesian Inference and Uncertainty Quantification for Inverse ProblemsMatt 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
 
bayesImageS: Bayesian computation for medical Image Segmentation using a hidd...
bayesImageS: Bayesian computation for medical Image Segmentation using a hidd...bayesImageS: Bayesian computation for medical Image Segmentation using a hidd...
bayesImageS: Bayesian computation for medical Image Segmentation using a hidd...Matt 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
 
R package 'bayesImageS': a case study in Bayesian computation using Rcpp and ...
R package 'bayesImageS': a case study in Bayesian computation using Rcpp and ...R package 'bayesImageS': a case study in Bayesian computation using Rcpp and ...
R package 'bayesImageS': a case study in Bayesian computation using Rcpp and ...Matt 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
 
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 (11)

Bayesian Inference and Uncertainty Quantification for Inverse Problems
Bayesian Inference and Uncertainty Quantification for Inverse ProblemsBayesian Inference and Uncertainty Quantification for Inverse Problems
Bayesian Inference and Uncertainty Quantification for Inverse Problems
 
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
 
bayesImageS: Bayesian computation for medical Image Segmentation using a hidd...
bayesImageS: Bayesian computation for medical Image Segmentation using a hidd...bayesImageS: Bayesian computation for medical Image Segmentation using a hidd...
bayesImageS: Bayesian computation for medical Image Segmentation using a hidd...
 
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
 
R package 'bayesImageS': a case study in Bayesian computation using Rcpp and ...
R package 'bayesImageS': a case study in Bayesian computation using Rcpp and ...R package 'bayesImageS': a case study in Bayesian computation using Rcpp and ...
R package 'bayesImageS': a case study in Bayesian computation using Rcpp and ...
 
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
 
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

GLYCOSIDES Classification Of GLYCOSIDES Chemical Tests Glycosides
GLYCOSIDES Classification Of GLYCOSIDES  Chemical Tests GlycosidesGLYCOSIDES Classification Of GLYCOSIDES  Chemical Tests Glycosides
GLYCOSIDES Classification Of GLYCOSIDES Chemical Tests GlycosidesNandakishor Bhaurao Deshmukh
 
Observational constraints on mergers creating magnetism in massive stars
Observational constraints on mergers creating magnetism in massive starsObservational constraints on mergers creating magnetism in massive stars
Observational constraints on mergers creating magnetism in massive starsSérgio Sacani
 
Quarter 4_Grade 8_Digestive System Structure and Functions
Quarter 4_Grade 8_Digestive System Structure and FunctionsQuarter 4_Grade 8_Digestive System Structure and Functions
Quarter 4_Grade 8_Digestive System Structure and FunctionsCharlene Llagas
 
Vision and reflection on Mining Software Repositories research in 2024
Vision and reflection on Mining Software Repositories research in 2024Vision and reflection on Mining Software Repositories research in 2024
Vision and reflection on Mining Software Repositories research in 2024AyushiRastogi48
 
STOPPED FLOW METHOD & APPLICATION MURUGAVENI B.pptx
STOPPED FLOW METHOD & APPLICATION MURUGAVENI B.pptxSTOPPED FLOW METHOD & APPLICATION MURUGAVENI B.pptx
STOPPED FLOW METHOD & APPLICATION MURUGAVENI B.pptxMurugaveni B
 
Biological classification of plants with detail
Biological classification of plants with detailBiological classification of plants with detail
Biological classification of plants with detailhaiderbaloch3
 
REVISTA DE BIOLOGIA E CIÊNCIAS DA TERRA ISSN 1519-5228 - Artigo_Bioterra_V24_...
REVISTA DE BIOLOGIA E CIÊNCIAS DA TERRA ISSN 1519-5228 - Artigo_Bioterra_V24_...REVISTA DE BIOLOGIA E CIÊNCIAS DA TERRA ISSN 1519-5228 - Artigo_Bioterra_V24_...
REVISTA DE BIOLOGIA E CIÊNCIAS DA TERRA ISSN 1519-5228 - Artigo_Bioterra_V24_...Universidade Federal de Sergipe - UFS
 
Citronella presentation SlideShare mani upadhyay
Citronella presentation SlideShare mani upadhyayCitronella presentation SlideShare mani upadhyay
Citronella presentation SlideShare mani upadhyayupadhyaymani499
 
Manassas R - Parkside Middle School 🌎🏫
Manassas R - Parkside Middle School 🌎🏫Manassas R - Parkside Middle School 🌎🏫
Manassas R - Parkside Middle School 🌎🏫qfactory1
 
The dark energy paradox leads to a new structure of spacetime.pptx
The dark energy paradox leads to a new structure of spacetime.pptxThe dark energy paradox leads to a new structure of spacetime.pptx
The dark energy paradox leads to a new structure of spacetime.pptxEran Akiva Sinbar
 
Base editing, prime editing, Cas13 & RNA editing and organelle base editing
Base editing, prime editing, Cas13 & RNA editing and organelle base editingBase editing, prime editing, Cas13 & RNA editing and organelle base editing
Base editing, prime editing, Cas13 & RNA editing and organelle base editingNetHelix
 
Davis plaque method.pptx recombinant DNA technology
Davis plaque method.pptx recombinant DNA technologyDavis plaque method.pptx recombinant DNA technology
Davis plaque method.pptx recombinant DNA technologycaarthichand2003
 
《Queensland毕业文凭-昆士兰大学毕业证成绩单》
《Queensland毕业文凭-昆士兰大学毕业证成绩单》《Queensland毕业文凭-昆士兰大学毕业证成绩单》
《Queensland毕业文凭-昆士兰大学毕业证成绩单》rnrncn29
 
Thermodynamics ,types of system,formulae ,gibbs free energy .pptx
Thermodynamics ,types of system,formulae ,gibbs free energy .pptxThermodynamics ,types of system,formulae ,gibbs free energy .pptx
Thermodynamics ,types of system,formulae ,gibbs free energy .pptxuniversity
 
User Guide: Capricorn FLX™ Weather Station
User Guide: Capricorn FLX™ Weather StationUser Guide: Capricorn FLX™ Weather Station
User Guide: Capricorn FLX™ Weather StationColumbia Weather Systems
 
CHROMATOGRAPHY PALLAVI RAWAT.pptx
CHROMATOGRAPHY  PALLAVI RAWAT.pptxCHROMATOGRAPHY  PALLAVI RAWAT.pptx
CHROMATOGRAPHY PALLAVI RAWAT.pptxpallavirawat456
 
Speech, hearing, noise, intelligibility.pptx
Speech, hearing, noise, intelligibility.pptxSpeech, hearing, noise, intelligibility.pptx
Speech, hearing, noise, intelligibility.pptxpriyankatabhane
 
GenAI talk for Young at Wageningen University & Research (WUR) March 2024
GenAI talk for Young at Wageningen University & Research (WUR) March 2024GenAI talk for Young at Wageningen University & Research (WUR) March 2024
GenAI talk for Young at Wageningen University & Research (WUR) March 2024Jene van der Heide
 
User Guide: Orion™ Weather Station (Columbia Weather Systems)
User Guide: Orion™ Weather Station (Columbia Weather Systems)User Guide: Orion™ Weather Station (Columbia Weather Systems)
User Guide: Orion™ Weather Station (Columbia Weather Systems)Columbia Weather Systems
 
Pests of castor_Binomics_Identification_Dr.UPR.pdf
Pests of castor_Binomics_Identification_Dr.UPR.pdfPests of castor_Binomics_Identification_Dr.UPR.pdf
Pests of castor_Binomics_Identification_Dr.UPR.pdfPirithiRaju
 

Recently uploaded (20)

GLYCOSIDES Classification Of GLYCOSIDES Chemical Tests Glycosides
GLYCOSIDES Classification Of GLYCOSIDES  Chemical Tests GlycosidesGLYCOSIDES Classification Of GLYCOSIDES  Chemical Tests Glycosides
GLYCOSIDES Classification Of GLYCOSIDES Chemical Tests Glycosides
 
Observational constraints on mergers creating magnetism in massive stars
Observational constraints on mergers creating magnetism in massive starsObservational constraints on mergers creating magnetism in massive stars
Observational constraints on mergers creating magnetism in massive stars
 
Quarter 4_Grade 8_Digestive System Structure and Functions
Quarter 4_Grade 8_Digestive System Structure and FunctionsQuarter 4_Grade 8_Digestive System Structure and Functions
Quarter 4_Grade 8_Digestive System Structure and Functions
 
Vision and reflection on Mining Software Repositories research in 2024
Vision and reflection on Mining Software Repositories research in 2024Vision and reflection on Mining Software Repositories research in 2024
Vision and reflection on Mining Software Repositories research in 2024
 
STOPPED FLOW METHOD & APPLICATION MURUGAVENI B.pptx
STOPPED FLOW METHOD & APPLICATION MURUGAVENI B.pptxSTOPPED FLOW METHOD & APPLICATION MURUGAVENI B.pptx
STOPPED FLOW METHOD & APPLICATION MURUGAVENI B.pptx
 
Biological classification of plants with detail
Biological classification of plants with detailBiological classification of plants with detail
Biological classification of plants with detail
 
REVISTA DE BIOLOGIA E CIÊNCIAS DA TERRA ISSN 1519-5228 - Artigo_Bioterra_V24_...
REVISTA DE BIOLOGIA E CIÊNCIAS DA TERRA ISSN 1519-5228 - Artigo_Bioterra_V24_...REVISTA DE BIOLOGIA E CIÊNCIAS DA TERRA ISSN 1519-5228 - Artigo_Bioterra_V24_...
REVISTA DE BIOLOGIA E CIÊNCIAS DA TERRA ISSN 1519-5228 - Artigo_Bioterra_V24_...
 
Citronella presentation SlideShare mani upadhyay
Citronella presentation SlideShare mani upadhyayCitronella presentation SlideShare mani upadhyay
Citronella presentation SlideShare mani upadhyay
 
Manassas R - Parkside Middle School 🌎🏫
Manassas R - Parkside Middle School 🌎🏫Manassas R - Parkside Middle School 🌎🏫
Manassas R - Parkside Middle School 🌎🏫
 
The dark energy paradox leads to a new structure of spacetime.pptx
The dark energy paradox leads to a new structure of spacetime.pptxThe dark energy paradox leads to a new structure of spacetime.pptx
The dark energy paradox leads to a new structure of spacetime.pptx
 
Base editing, prime editing, Cas13 & RNA editing and organelle base editing
Base editing, prime editing, Cas13 & RNA editing and organelle base editingBase editing, prime editing, Cas13 & RNA editing and organelle base editing
Base editing, prime editing, Cas13 & RNA editing and organelle base editing
 
Davis plaque method.pptx recombinant DNA technology
Davis plaque method.pptx recombinant DNA technologyDavis plaque method.pptx recombinant DNA technology
Davis plaque method.pptx recombinant DNA technology
 
《Queensland毕业文凭-昆士兰大学毕业证成绩单》
《Queensland毕业文凭-昆士兰大学毕业证成绩单》《Queensland毕业文凭-昆士兰大学毕业证成绩单》
《Queensland毕业文凭-昆士兰大学毕业证成绩单》
 
Thermodynamics ,types of system,formulae ,gibbs free energy .pptx
Thermodynamics ,types of system,formulae ,gibbs free energy .pptxThermodynamics ,types of system,formulae ,gibbs free energy .pptx
Thermodynamics ,types of system,formulae ,gibbs free energy .pptx
 
User Guide: Capricorn FLX™ Weather Station
User Guide: Capricorn FLX™ Weather StationUser Guide: Capricorn FLX™ Weather Station
User Guide: Capricorn FLX™ Weather Station
 
CHROMATOGRAPHY PALLAVI RAWAT.pptx
CHROMATOGRAPHY  PALLAVI RAWAT.pptxCHROMATOGRAPHY  PALLAVI RAWAT.pptx
CHROMATOGRAPHY PALLAVI RAWAT.pptx
 
Speech, hearing, noise, intelligibility.pptx
Speech, hearing, noise, intelligibility.pptxSpeech, hearing, noise, intelligibility.pptx
Speech, hearing, noise, intelligibility.pptx
 
GenAI talk for Young at Wageningen University & Research (WUR) March 2024
GenAI talk for Young at Wageningen University & Research (WUR) March 2024GenAI talk for Young at Wageningen University & Research (WUR) March 2024
GenAI talk for Young at Wageningen University & Research (WUR) March 2024
 
User Guide: Orion™ Weather Station (Columbia Weather Systems)
User Guide: Orion™ Weather Station (Columbia Weather Systems)User Guide: Orion™ Weather Station (Columbia Weather Systems)
User Guide: Orion™ Weather Station (Columbia Weather Systems)
 
Pests of castor_Binomics_Identification_Dr.UPR.pdf
Pests of castor_Binomics_Identification_Dr.UPR.pdfPests of castor_Binomics_Identification_Dr.UPR.pdf
Pests of castor_Binomics_Identification_Dr.UPR.pdf
 

Approximate Bayesian computation for the Ising/Potts model

  • 1. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion an introduction to Approximate Bayesian Computation Matt Moores Research Fellow Department of Statistics University of Warwick Warwick ML Club June 12, 2017
  • 2. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion Motivation Inference for a parameter θ when it is: impossible or very expensive to evaluate the likelihood p(y|θ) ABC is a likelihood-free method for approximating the posterior distribution π(θ|y) by generating pseudo-data from the model: w ∼ f(·|θ)
  • 3. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion Likelihood-free rejection sampler Algorithm 1 Likelihood-free rejection sampler 1: Draw parameter value θ ∼ π(θ) 2: Generate w ∼ f(·|θ ) 3: if w = y (the observed data) then 4: accept θ 5: end if But if the observations y are continuous (or the space y ∈ Y is enormous) then P(w = y) ≈ 0 Tavar´e, Balding, Griffith & Donnelly (1997) Genetics 145(2)
  • 4. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion ABC tolerance accept θ if δ(w, y) < where > 0 is the tolerance level δ(·, ·) is a distance function (for an appropriate choice of norm) Inference is more exact when is close to zero. but more proposed θ are rejected (tradeoff between accuracy & computational cost) Pritchard, Seielstad, Perez-Lezaun & Feldman (1999) Mol. Biol. Evol. 16(12)
  • 5. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion Summary statistics Computing δ(w, y) for w1, . . . , wn and y1, . . . , yn can be very expensive for large n Instead, compute summary statistics s(y) e.g. sufficient statistics (only available for exponential family)
  • 6. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion Sufficient statistics Fisher-Neyman factorisation theorem: if s(y) is sufficient for θ then p(y|θ) = f(y) g (s(y)|θ) only applies to Potts, Ising, exponential random graph models (ERGM) otherwise, selection of suitable summary statistics can be a very difficult problem
  • 7. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion ABC rejection sampler Algorithm 2 ABC rejection sampler 1: for all iterations t ∈ 1 . . . T do 2: Draw independent proposal θ ∼ π(θ) 3: Generate w ∼ f(·|θ ) 4: if s(w) − s(y) < then 5: set θt ← θ 6: else 7: set θt ← θt−1 8: end if 9: end for Approximates π(θ|y) by π (θ | s(w) − s(y) < ) Marin, Pudlo, Robert & Ryder (2012) Stat. Comput. 22(6) Marin & Robert (2014) Bayesian Essentials with R §8.3
  • 8. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion A trivial example Gaussian with unknown variance: y ∼ N(1, σ2 ) normalising constant: Z(σ2 ) = (2πσ2 )−n 2 natural conjugate prior: π 1 σ2 ∼ Ga ν0 2 , ν0ψ2 0 2 sufficient statistic: s(y) = 1 n n i=1 (yi − 1)2 posterior is analytically tractable: π 1 σ2 | y ∼ Ga ν0+n 2 , ν0ψ2 0+ns(y) 2 ∴ no need for ABC (nor MCMC) in practice
  • 9. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion R code π(τ|y) Density 0 1 2 3 4 5 6 7 0.00.10.20.30.4 § y ← rnorm (n=5, mean=1, sd=2/3) n ← length ( y ) s sq ← sum(( y −1)ˆ2)/n post nu ← nu0 + n post ssd ← ( nu0 ∗ s0 ˆ2 + n∗ s sq )/2
  • 10. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion now with ABC π(τ) 0 1 2 3 4 5 6 7 0.00.51.01.5 πε(τ | δ(s(w), s(y)) < ε) 0 1 2 3 4 5 6 7 0.00.10.20.30.4 § prop tau ← rgamma(10000 , nu0/2 , 0.5 ∗nu0∗ s0 ˆ2) pseudo ← rnorm (n∗ 10000 , 1 , 1/ sqrt ( prop tau )) pseudoMx ← matrix ( pseudo , nrow=10000, ncol=n) pseudoSSD ← rowSums (( pseudoMx − 1)ˆ2)/n ps norm ← abs ( pseudoSSD − s sq ) e p s i l o n ← sort ( ps norm ) [ 2 0 0 0 ] prop keep ← prop tau [ ps norm <= e p s i l o n ]
  • 11. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion choice of 0 1 2 3 4 5 6 7 0.00.20.40.6 (a) = 8.430 0 1 2 3 4 5 6 7 0.00.10.20.30.4 (b) = 1.427 0 1 2 3 4 5 6 7 0.00.20.40.6 (c) = 0.011 0 1 2 3 4 5 6 7 0.00.40.8 (d) = 0.001
  • 12. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion concentration of measure 0 1 2 3 4 0.00.40.8 (a) n = 25, = 0.149 0 1 2 3 4 0.01.02.03.0 (b) n = 500, = 0.025 0 1 2 3 4 02468 (c) n = 104 , = 0.011 0 1 2 3 4 020406080100 (d) n = 106 , = 0.001
  • 13. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion Metropolis-Hastings proposals Algorithm 3 ABC-MCMC 1: Initialise θ0 ∼ π(θ) 2: for all iterations t ∈ 1 . . . T do 3: Draw proposal θ ∼ q(· | θt−1) 4: Generate w ∼ f(·|θ ) 5: Draw u ∼ Unif(0, 1) 6: if u < π(θ )q(θt−1|θ ) π(θt−1)q(θ |θt−1) and s(w) − s(y) < then 7: set θt ← θ 8: else 9: set θt ← θt−1 10: end if 11: end for Unfortunately, this algorithm is prone to getting ”stuck” Marjoram, Molitor, Plagnol & Tavar´e (2003) PNAS 100(26) Lee & Latuszy´nski (2014) Biometrika 101(3)
  • 14. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion Sequential Monte Carlo Algorithm 4 ABC-SMC 1: Draw N particles θi ∼ π0(θ) 2: Draw N × M sets of pseudo-data wi,m ∼ f(·|θi) 3: repeat 4: Adaptively select ABC tolerance t 5: Update importance weights λi for each particle 6: if effective sample size (ESS) < Nmin then 7: Resample particles according to their weights 8: end if 9: Update particles using Metropolis-Hastings step (with adaptive proposal bandwidth σ2 t ) 10: until naccept N < 0.015 or t < 10−9 or t ≥ 100 Targets a sequence of distributions π t (θ | s(w) − s(y) < t) such that 1 > 2 > · · · > T Drovandi & Pettitt (2011) Biometrics 67(1) Del Moral, Doucet & Jasra (2012) Stat. Comput. 22(5)
  • 15. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion hidden Markov random field Joint distribution of observed pixel intensities y = (y1, . . . , yn) ∈ Rn and latent labels z = (z1, . . . , zn) ∈ {1, . . . , k}n: Pr(y, z | µ, σ2 , β) = L(y|µ, σ2 , z)π(z|β) (1) Additive Gaussian noise: yi | zi =j iid ∼ N µj, σ2 j (2) Potts model: π(zi | zi, β) = exp {β i∼ δ(zi, z )} k j=1 exp {β i∼ δ(j, z )} (3) Potts (1952) Proceedings of the Cambridge Philosophical Society 48(1) Winkler (2003) Image Analysis, Random Fields and MCMC Methods, 2nd ed.
  • 16. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion Inverse Temperature (e) β = 0.1 (f) β = 0.5 (g) β = 0.85 (h) β = 0.95 (i) β = 1.005 (j) β = 1.15
  • 17. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion Doubly-intractable posterior p(β|z) = C−1(β)eβS(z)π(β) β C−1(β)eβS(z)π(dβ) (4) The normalising constant has computational complexity O(nkn), since it involves a sum over all possible combinations of the labels z ∈ Z: C(β) = z∈Z eβS(z) (5) S(z) is the sufficient statistic of the Potts model: S(z) = i∼ ∈E δ(zi, z ) (6) where E is the set of all unique neighbour pairs.
  • 18. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion bayesImageS An R package for Bayesian image segmentation using the hidden Potts model: RcppArmadillo for fast computation in C++ OpenMP for parallelism § l i b r a r y ( bayesImageS ) p r i o r s ← l i s t ("k"=3,"mu"=rep (0 ,3) , "mu.sd"=sigma , "sigma"=sigma , "sigma.nu"=c (1 ,1 ,1) , "beta"=c (0 ,3)) mh ← l i s t ( algorithm="pseudo" , bandwidth =0.2) r e s u l t ← mcmcPotts ( y , neigh , block ,NULL, 55000 ,5000 , p r i o r s ,mh) Eddelbuettel & Sanderson (2014) RcppArmadillo: Accelerating R with high-performance C++ linear algebra. CSDA 71
  • 19. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion Bayesian computational methods bayesImageS supports methods for updating the latent labels z: Chequerboard updating (Winkler 2003) Swendsen-Wang (1987) and also methods for updating the inverse temperature β: Pseudolikelihood (Ryd´en & Titterington 1998) Path Sampling (Gelman & Meng 1998) Exchange Algorithm (Murray, Ghahramani & MacKay 2006) Approximate Bayesian Computation (Grelaud et al. 2009) Sequential Monte Carlo (SMC-ABC) (Del Moral, Doucet & Jasra 2012)
  • 20. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion Lake Menteith, Scotland 0.0 0.2 0.4 0.6 0.8 1.0 0.00.20.40.60.81.0
  • 21. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion ABC-SMC for the hidden Potts model § l i b r a r y ( bayess ) data ( Menteith ) y ← as . matrix ( Menteith ) l i b r a r y ( bayesImageS ) mask ← matrix (1 , nrow=nrow( y ) , ncol=ncol ( y )) neigh ← getNeighbors (mask , c (2 ,2 ,0 ,0)) block ← getBlocks (mask , 2) p r i o r s ← l i s t ( k=6, mu=rep (256/2 , 6) , mu. sd=rep (256/ 6 ,6) , sigma=rep (256/ 6 ,6) , sigma . nu=rep (6 , 6) , beta ← c (0 ,2)) r e s ← smcPotts ( as . vector ( y ) , neigh , block , param=l i s t ( npart =2000, nstat =5) , p r i o r s=p r i o r s ) 6h 15min for 100 SMC iterations (N=2000, M=5)
  • 22. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion ABC tolerance SMC iteration εt 0 20 40 60 80 100 02000400060008000100001200014000
  • 23. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion approximate posterior ( t = 37.5) β Density 1.280 1.285 1.290 1.295 1.300 1.305 1.310 1.315 0204060 Equivalent to 120,801 iterations of the ABC rejection sampler
  • 24. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion Effective Sample Size SMC iteration ESS 0 20 40 60 80 100 800100012001400160018002000
  • 25. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion image segmentation (a) Original image (b) Potts labels
  • 26. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion Summary ABC is a method for likelihood-free inference It enables inference for models that are otherwise computationally intractable Main components of ABC: π(θ) proposal density for θ f(·|θ) generative model for w tolerance level δ(·, ·) distance function s(y) summary statistics
  • 27. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion References I M. Moores, A. N. Pettitt & K. Mengersen Scalable Bayesian inference for the inverse temperature of a hidden Potts model. arXiv:1503.08066 [stat.CO], 2015. M. Moores, C. C. Drovandi, K. Mengersen & C. P. Robert Pre-processing for approximate Bayesian computation in image analysis. Statistics & Computing 25(1): 23–33, 2015. J.-M. Marin, P. Pudlo, C. P. Robert & R. Ryder Approximate Bayesian computational methods. Statistics & Computing, 22(6): 1167–80, 2012. A. Grelaud, C. P. Robert, J.-M. Marin, F. Rodolphe & J.-F. Taly ABC likelihood-free methods for model choice in Gibbs random fields. Bayesian Analysis, 4(2): 317–36, 2009. J.-M. Marin & C. P. Robert Bayesian Essentials with R Springer-Verlag, 2014.
  • 28. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion References II A. Lee & K. Latuszy´nski Variance bounding and geometric ergodicity of Markov chain Monte Carlo kernels for approximate Bayesian computation Biometrika 101(3): 655–671, 2014. P. Del Moral, A. Doucet & A. Jasra An adaptive sequential Monte Carlo method for approximate Bayesian computation. Statistics & Computing, 22(5): 1009–20, 2012. C. C. Drovandi & A. N. Pettitt Estimation of Parameters for Macroparasite Population Evolution Using Approximate Bayesian Computation Biometrics 67(1): 225–233, 2011. P. Marjoram, J. Molitor, V. Plagnol & S. Tavar´e Markov chain Monte Carlo without likelihoods. Proc. Natl Acad. Sci. USA, 100(26): 15324–15328, 2003.
  • 29. Intro to ABC Simulation Study ABC Algorithms Ising/Potts model Image Analysis Conclusion References III J. Pritchard, M. Seielstad, A. Perez-Lezaun & M. Feldman Population Growth of Human Y Chromosomes: A Study of Y Chromosome Microsatellites. Mol. Biol. Evol. 16(12): 1791–98, 1999. S. Tavar´e, D. Balding, R, Griffiths & P. Donnelly Inferring coalescence times from DNA sequence data. Genetics, 145(2): 505–18, 1997. R. B. Potts Some generalized order-disorder transformations. Proc. Cambridge Philosophical Society, 48(1): 106–109, 1952. G. Winkler Image analysis, random fields and Markov chain Monte Carlo methods 2nd ed., Springer-Verlag, 2003. D. Eddelbuettel Seamless R and C++ integration with Rcpp Springer-Verlag, 2013.