SlideShare uma empresa Scribd logo
1 de 63
Baixar para ler offline
ABC Methods for Bayesian Model Choice




                 ABC Methods for Bayesian Model Choice

                                        Christian P. Robert

                                Universit´ Paris-Dauphine, IuF, & CREST
                                         e
                               http://www.ceremade.dauphine.fr/~xian


                                           May 4, 2011
ABC Methods for Bayesian Model Choice




ABC for model choice



      Model choice

      Gibbs random fields

      Generic ABC model choice
ABC Methods for Bayesian Model Choice
  Model choice




Bayesian model choice



      Several models M1 , M2 , . . . are considered simultaneously for a
      dataset y and the model index M is part of the inference.
      Use of a prior distribution. π(M = m), plus a prior distribution on
      the parameter conditional on the value m of the model index,
      πm (θ m )
      Goal is to derive the posterior distribution of M , challenging
      computational target when models are complex.
ABC Methods for Bayesian Model Choice
  Model choice




Generic ABC for model choice


      Algorithm 1 Likelihood-free model choice sampler (ABC-MC)
        for t = 1 to T do
          repeat
             Generate m from the prior π(M = m)
             Generate θ m from the prior πm (θ m )
             Generate z from the model fm (z|θ m )
          until ρ{η(z), η(y)} <
          Set m(t) = m and θ (t) = θ m
        end for

                                   [Toni, Welch, Strelkowa, Ipsen & Stumpf, 2009]
ABC Methods for Bayesian Model Choice
  Model choice




ABC estimates
      Posterior probability π(M = m|y) approximated by the frequency
      of acceptances from model m
                                             T
                                         1
                                                   Im(t) =m .
                                         T
                                             t=1

      Early issues with implementation:
             should tolerances          be the same for all models?
             should summary statistics vary across models? (incl. their
             dimension)?
             should the distance measure ρ vary as well?
ABC Methods for Bayesian Model Choice
  Model choice




ABC estimates

      Posterior probability π(M = m|y) approximated by the frequency
      of acceptances from model m
                                            T
                                        1
                                                  Im(t) =m .
                                        T
                                            t=1

      Early issues with implementation:
                 then needs to become part of the model
             this is incompatible with Bayesian model choice
             same difficulty
      Extension to a weighted polychotomous logistic regression estimate
      of π(M = m|y), with non-parametric kernel weights
                                        [Cornuet et al., DIYABC, 2009]
ABC Methods for Bayesian Model Choice
  Model choice




The Great ABC controversy
      On-going controvery in phylogeographic genetics about the validity
      of using ABC for testing




   Against: Templeton, 2008,
   2009, 2010a, 2010b, 2010c
   argues that nested hypotheses
   cannot have higher probabilities
   than nesting hypotheses (!)
ABC Methods for Bayesian Model Choice
  Model choice




The Great ABC controversy



     On-going controvery in phylogeographic genetics about the validity
     of using ABC for testing
                                        Replies: Fagundes et al., 2008,
   Against: Templeton, 2008,           Beaumont et al., 2010, Berger et
   2009, 2010a, 2010b, 2010c           al., 2010, Csill`ry et al., 2010
                                                       e
   argues that nested hypotheses       point out that the criticisms are
   cannot have higher probabilities    addressed at [Bayesian]
   than nesting hypotheses (!)         model-based inference and have
                                       nothing to do with ABC...
ABC Methods for Bayesian Model Choice
  Gibbs random fields




Gibbs random fields


      Gibbs distribution
      The rv y = (y1 , . . . , yn ) is a Gibbs random field associated with
      the graph G if

                                           1
                                 f (y) =     exp −         Vc (yc )   ,
                                           Z
                                                     c∈C

      where Z is the normalising constant, C is the set of cliques of G
      and Vc is any function also called potential
      U (y) = c∈C Vc (yc ) is the energy function
ABC Methods for Bayesian Model Choice
  Gibbs random fields




Gibbs random fields


      Gibbs distribution
      The rv y = (y1 , . . . , yn ) is a Gibbs random field associated with
      the graph G if

                                           1
                                 f (y) =     exp −         Vc (yc )   ,
                                           Z
                                                     c∈C

      where Z is the normalising constant, C is the set of cliques of G
      and Vc is any function also called potential
      U (y) = c∈C Vc (yc ) is the energy function

       c Z is usually unavailable in closed form
ABC Methods for Bayesian Model Choice
  Gibbs random fields




Potts model

      Potts model
      Vc (y) is of the form

                                   Vc (y) = θS(y) = θ         δyl =yi
                                                        l∼i

      where l∼i denotes a neighbourhood structure
ABC Methods for Bayesian Model Choice
  Gibbs random fields




Potts model

      Potts model
      Vc (y) is of the form

                                   Vc (y) = θS(y) = θ            δyl =yi
                                                           l∼i

      where l∼i denotes a neighbourhood structure

      In most realistic settings, summation

                                        Zθ =         exp{θ T S(x)}
                                               x∈X

      involves too many terms to be manageable and numerical
      approximations cannot always be trusted
ABC Methods for Bayesian Model Choice
  Gibbs random fields




Bayesian Model Choice



      Comparing a model with energy S0 taking values in Rp0 versus a
      model with energy S1 taking values in Rp1 can be done through
      the Bayes factor corresponding to the priors π0 and π1 on each
      parameter space

                                        exp{θ T S0 (x)}/Zθ 0 ,0 π0 (dθ 0 )
                                              0
                       Bm0 /m1 (x) =
                                        exp{θ T S1 (x)}/Zθ 1 ,1 π1 (dθ 1 )
                                              1
ABC Methods for Bayesian Model Choice
  Gibbs random fields




Neighbourhood relations



      Choice to be made between M neighbourhood relations
                                        m
                                    i∼i           (0 ≤ m ≤ M − 1)

      with
                                            Sm (x) =         I{xi =xi }
                                                       m
                                                       i∼i

      driven by the posterior probabilities of the models.
ABC Methods for Bayesian Model Choice
  Gibbs random fields




Model index


      Computational target:

              P(M = m|x) ∝                   fm (x|θm )πm (θm ) dθm π(M = m) ,
                                        Θm
ABC Methods for Bayesian Model Choice
  Gibbs random fields




Model index


      Computational target:

              P(M = m|x) ∝                   fm (x|θm )πm (θm ) dθm π(M = m) ,
                                        Θm

      If S(x) sufficient statistic for the joint parameters
      (M, θ0 , . . . , θM −1 ),

                               P(M = m|x) = P(M = m|S(x)) .
ABC Methods for Bayesian Model Choice
  Gibbs random fields




Sufficient statistics in Gibbs random fields
ABC Methods for Bayesian Model Choice
  Gibbs random fields




Sufficient statistics in Gibbs random fields


      Each model m has its own sufficient statistic Sm (·) and
      S(·) = (S0 (·), . . . , SM −1 (·)) is also (model-)sufficient.
ABC Methods for Bayesian Model Choice
  Gibbs random fields




Sufficient statistics in Gibbs random fields


      Each model m has its own sufficient statistic Sm (·) and
      S(·) = (S0 (·), . . . , SM −1 (·)) is also (model-)sufficient.
      For Gibbs random fields,
                                        1           2
                x|M = m ∼ fm (x|θm ) = fm (x|S(x))fm (S(x)|θm )
                                           1
                                     =         f 2 (S(x)|θm )
                                       n(S(x)) m

      where
                              n(S(x)) = {˜ ∈ X : S(˜ ) = S(x)}
                                         x         x
       c S(x) is therefore also sufficient for the joint parameters
ABC Methods for Bayesian Model Choice
  Gibbs random fields




ABC model choice Algorithm



      ABC-MC
             Generate m∗ from the prior π(M = m).
                       ∗
             Generate θm∗ from the prior πm∗ (·).
             Generate x∗ from the model fm∗ (·|θm∗ ).
                                                ∗

             Compute the distance ρ(S(x0 ), S(x∗ )).
             Accept (θm∗ , m∗ ) if ρ(S(x0 ), S(x∗ )) < .
                      ∗


      Note When            = 0 the algorithm is exact
ABC Methods for Bayesian Model Choice
  Gibbs random fields




Toy example


      iid Bernoulli model versus two-state first-order Markov chain, i.e.
                                              n
                  f0 (x|θ0 ) = exp θ0             I{xi =1}     {1 + exp(θ0 )}n ,
                                           i=1

      versus
                                          n
                               1
             f1 (x|θ1 ) =        exp θ1         I{xi =xi−1 }    {1 + exp(θ1 )}n−1 ,
                               2
                                          i=2

      with priors θ0 ∼ U(−5, 5) and θ1 ∼ U(0, 6) (inspired by “phase
      transition” boundaries).
ABC Methods for Bayesian Model Choice
  Gibbs random fields




Toy example (2)
                                                                 ^
                                                                BF01

                                                      −5        0           5




                                          −40
                                          −20
                                   BF01

                                          0
                                          10




                                                                 ^
                                                                BF01

                                                −10        −5       0   5       10
                                          −40
                                          −20
                                   BF01

                                          0
                                          10




      (left) Comparison of the true BF m0 /m1 (x0 ) with BF m0 /m1 (x0 )
      (in logs) over 2, 000 simulations and 4.106 proposals from the
      prior. (right) Same when using tolerance corresponding to the
      1% quantile on the distances.
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




Back to sufficiency


             ‘Sufficient statistics for individual models are unlikely to
             be very informative for the model probability. This is
             already well known and understood by the ABC-user
             community.’
                                          [Scott Sisson, Jan. 31, 2011, ’Og]
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




Back to sufficiency


             ‘Sufficient statistics for individual models are unlikely to
             be very informative for the model probability. This is
             already well known and understood by the ABC-user
             community.’
                                          [Scott Sisson, Jan. 31, 2011, ’Og]

      If η1 (x) sufficient statistic for model m = 1 and parameter θ1 and
      η2 (x) sufficient statistic for model m = 2 and parameter θ2 ,
      (η1 (x), η2 (x)) is not always sufficient for (m, θm )
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




Back to sufficiency


             ‘Sufficient statistics for individual models are unlikely to
             be very informative for the model probability. This is
             already well known and understood by the ABC-user
             community.’
                                          [Scott Sisson, Jan. 31, 2011, ’Og]

      If η1 (x) sufficient statistic for model m = 1 and parameter θ1 and
      η2 (x) sufficient statistic for model m = 2 and parameter θ2 ,
      (η1 (x), η2 (x)) is not always sufficient for (m, θm )
       c Potential loss of information at the testing level
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




Limiting behaviour of B12 (T → ∞)

      ABC approximation
                                         T
                                         t=1 Imt =1 Iρ{η(zt ),η(y)}≤
                             B12 (y) =   T
                                                                       ,
                                         t=1 Imt =2 Iρ{η(zt ),η(y)}≤

      where the (mt , z t )’s are simulated from the (joint) prior
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




Limiting behaviour of B12 (T → ∞)

      ABC approximation
                                            T
                                            t=1 Imt =1 Iρ{η(zt ),η(y)}≤
                             B12 (y) =      T
                                                                          ,
                                            t=1 Imt =2 Iρ{η(zt ),η(y)}≤

      where the (mt , z t )’s are simulated from the (joint) prior
      As T go to infinity, limit

                                        Iρ{η(z),η(y)}≤ π1 (θ 1 )f1 (z|θ 1 ) dz dθ 1
                  B12 (y) =
                                        Iρ{η(z),η(y)}≤ π2 (θ 2 )f2 (z|θ 2 ) dz dθ 2
                                                              η
                                        Iρ{η,η(y)}≤ π1 (θ 1 )f1 (η|θ 1 ) dη dθ 1
                                =                             η                  ,
                                        Iρ{η,η(y)}≤ π2 (θ 2 )f2 (η|θ 2 ) dη dθ 2
             η               η
      where f1 (η|θ 1 ) and f2 (η|θ 2 ) distributions of η(z)
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




Limiting behaviour of B12 ( → 0)




      When         goes to zero,
                                                   η
                              η          π1 (θ 1 )f1 (η(y)|θ 1 ) dθ 1
                             B12 (y) =             η                  ,
                                         π2 (θ 2 )f2 (η(y)|θ 2 ) dθ 2
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




Limiting behaviour of B12 ( → 0)




      When         goes to zero,
                                                   η
                              η          π1 (θ 1 )f1 (η(y)|θ 1 ) dθ 1
                             B12 (y) =             η                  ,
                                         π2 (θ 2 )f2 (η(y)|θ 2 ) dθ 2

                  Bayes factor based on the sole observation of η(y)
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




Limiting behaviour of B12 (under sufficiency)
      If η(y) sufficient statistic for both models,

                                    fi (y|θ i ) = gi (y)fiη (η(y)|θ i )

      Thus
                                                                 η
                                              Θ1   π(θ 1 )g1 (y)f1 (η(y)|θ 1 ) dθ 1
                       B12 (y) =                                 η
                                              Θ2   π(θ 2 )g2 (y)f2 (η(y)|θ 2 ) dθ 2
                                                              η
                                             g1 (y) π1 (θ 1 )f1 (η(y)|θ 1 ) dθ 1
                                        =                     η
                                             g2 (y) π2 (θ 2 )f2 (η(y)|θ 2 ) dθ 2
                                             g1 (y) η
                                        =          B (y) .
                                             g2 (y) 12

                                            [Didelot, Everitt, Johansen & Lawson, 2011]
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




Limiting behaviour of B12 (under sufficiency)
      If η(y) sufficient statistic for both models,

                                    fi (y|θ i ) = gi (y)fiη (η(y)|θ i )

      Thus
                                                                 η
                                              Θ1   π(θ 1 )g1 (y)f1 (η(y)|θ 1 ) dθ 1
                       B12 (y) =                                 η
                                              Θ2   π(θ 2 )g2 (y)f2 (η(y)|θ 2 ) dθ 2
                                                              η
                                             g1 (y) π1 (θ 1 )f1 (η(y)|θ 1 ) dθ 1
                                        =                     η
                                             g2 (y) π2 (θ 2 )f2 (η(y)|θ 2 ) dθ 2
                                             g1 (y) η
                                        =          B (y) .
                                             g2 (y) 12

                                            [Didelot, Everitt, Johansen & Lawson, 2011]
                    No discrepancy only when cross-model sufficiency
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




Poisson/geometric example

      Sample
                                           x = (x1 , . . . , xn )
      from either a Poisson P(λ) or from a geometric G(p)
      Sum
                                                 n
                                          S=          yi = η(x)
                                                i=1

      sufficient statistic for either model but not simultaneously
      Discrepancy ratio

                                        g1 (x)   S!n−S / i yi !
                                               =
                                        g2 (x)    1 n+S−1
                                                        S
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




Poisson/geometric discrepancy

                               η
      Range of B12 (x) versus B12 (x) B12 (x): The values produced have
      nothing in common.
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




Formal recovery



      Creating an encompassing exponential family
                                      T           T
      f (x|θ1 , θ2 , α1 , α2 ) ∝ exp{θ1 η1 (x) + θ1 η1 (x) + α1 t1 (x) + α2 t2 (x)}

      leads to a sufficient statistic (η1 (x), η2 (x), t1 (x), t2 (x))
                             [Didelot, Everitt, Johansen & Lawson, 2011]
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




Formal recovery



      Creating an encompassing exponential family
                                      T           T
      f (x|θ1 , θ2 , α1 , α2 ) ∝ exp{θ1 η1 (x) + θ1 η1 (x) + α1 t1 (x) + α2 t2 (x)}

      leads to a sufficient statistic (η1 (x), η2 (x), t1 (x), t2 (x))
                             [Didelot, Everitt, Johansen & Lawson, 2011]

      In the Poisson/geometric case, if        i xi !   is added to S, no
      discrepancy
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




Formal recovery



      Creating an encompassing exponential family
                                      T           T
      f (x|θ1 , θ2 , α1 , α2 ) ∝ exp{θ1 η1 (x) + θ1 η1 (x) + α1 t1 (x) + α2 t2 (x)}

      leads to a sufficient statistic (η1 (x), η2 (x), t1 (x), t2 (x))
                             [Didelot, Everitt, Johansen & Lawson, 2011]

      Only applies in genuine sufficiency settings...
      c Inability to evaluate loss brought by summary statistics
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




The Pitman–Koopman lemma



      Lemma
      A necessary and sufficient condition for the existence of a sufficient
      statistic of a fixed dimension whatever the sample size is that the
      sampling distribution belongs to an exponential family.

                                        [Pitman, 1933; Koopman, 1933]
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




The Pitman–Koopman lemma



      Lemma
      A necessary and sufficient condition for the existence of a sufficient
      statistic of a fixed dimension whatever the sample size is that the
      sampling distribution belongs to an exponential family.

                                        [Pitman, 1933; Koopman, 1933]

      Provision of fixed support (consider U(0, θ) counterexample)
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




Meaning of the ABC-Bayes factor

             ‘This is also why focus on model discrimination typically
             (...) proceeds by (...) accepting that the Bayes Factor
             that one obtains is only derived from the summary
             statistics and may in no way correspond to that of the
             full model.’
                                         [Scott Sisson, Jan. 31, 2011, ’Og]
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




Meaning of the ABC-Bayes factor

             ‘This is also why focus on model discrimination typically
             (...) proceeds by (...) accepting that the Bayes Factor
             that one obtains is only derived from the summary
             statistics and may in no way correspond to that of the
             full model.’
                                                [Scott Sisson, Jan. 31, 2011, ’Og]

      In the Poisson/geometric case, if E[yi ] = θ0 > 0,

                                         η          (θ0 + 1)2 −θ0
                                    lim B12 (y) =            e
                                   n→∞                  θ0
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




MA example



      Consider the MA(q) model
                                                     q
                                        xt =   t+         ϑi   t−i
                                                    i=1

      Simple prior: uniform prior over the identifiability zone, e.g.
      triangle for MA(2)
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




MA example (2)
      ABC algorithm thus made of
         1. picking a new value (ϑ1 , ϑ2 ) in the triangle
         2. generating an iid sequence ( t )−q<t≤T
         3. producing a simulated series (xt )1≤t≤T
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




MA example (2)
      ABC algorithm thus made of
         1. picking a new value (ϑ1 , ϑ2 ) in the triangle
         2. generating an iid sequence ( t )−q<t≤T
         3. producing a simulated series (xt )1≤t≤T
      Distance: basic distance between the series
                                                              T
                             ρ((xt )1≤t≤T , (xt )1≤t≤T ) =         (xt − xt )2
                                                             t=1

      or between summary statistics like the first q autocorrelations
                                                 T
                                         τj =           xt xt−j
                                                t=j+1
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




Comparison of distance impact




      Evaluation of the tolerance on the ABC sample against both
      distances ( = 100%, 10%, 1%, 0.1%) for an MA(2) model
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




Comparison of distance impact

                      4




                                                           1.5
                      3




                                                           1.0
                      2




                                                           0.5
                      1




                                                           0.0
                      0




                             0.0   0.2   0.4   0.6   0.8         −2.0   −1.0    0.0   0.5   1.0   1.5

                                          θ1                                   θ2




      Evaluation of the tolerance on the ABC sample against both
      distances ( = 100%, 10%, 1%, 0.1%) for an MA(2) model
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




Comparison of distance impact

                      4




                                                           1.5
                      3




                                                           1.0
                      2




                                                           0.5
                      1




                                                           0.0
                      0




                             0.0   0.2   0.4   0.6   0.8         −2.0   −1.0    0.0   0.5   1.0   1.5

                                          θ1                                   θ2




      Evaluation of the tolerance on the ABC sample against both
      distances ( = 100%, 10%, 1%, 0.1%) for an MA(2) model
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




MA(q) divergence




                                        1.0




                                                      1.0




                                                                    1.0




                                                                                  1.0
                                        0.8




                                                      0.8




                                                                    0.8




                                                                                  0.8
                                        0.6




                                                      0.6




                                                                    0.6




                                                                                  0.6
                                        0.4




                                                      0.4




                                                                    0.4




                                                                                  0.4
                                        0.2




                                                      0.2




                                                                    0.2




                                                                                  0.2
                                        0.0




                                                      0.0




                                                                    0.0




                                                                                  0.0
                                              1   2         1   2         1   2         1   2




      Evolution [against ] of ABC Bayes factor, in terms of frequencies of
      visits to models MA(1) (left) and MA(2) (right) when equal to
      10, 1, .1, .01% quantiles on insufficient autocovariance distances. Sample
      of 50 points from a MA(2) with θ1 = 0.6, θ2 = 0.2. True Bayes factor
      equal to 17.71.
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




MA(q) divergence




                                        1.0




                                                      1.0




                                                                    1.0




                                                                                  1.0
                                        0.8




                                                      0.8




                                                                    0.8




                                                                                  0.8
                                        0.6




                                                      0.6




                                                                    0.6




                                                                                  0.6
                                        0.4




                                                      0.4




                                                                    0.4




                                                                                  0.4
                                        0.2




                                                      0.2




                                                                    0.2




                                                                                  0.2
                                        0.0




                                                      0.0




                                                                    0.0




                                                                                  0.0
                                              1   2         1   2         1   2         1   2




      Evolution [against ] of ABC Bayes factor, in terms of frequencies of
      visits to models MA(1) (left) and MA(2) (right) when equal to
      10, 1, .1, .01% quantiles on insufficient autocovariance distances. Sample
      of 50 points from a MA(1) model with θ1 = 0.6. True Bayes factor B21
      equal to .004.
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




Further comments


             ‘There should be the possibility that for the same model,
             but different (non-minimal) [summary] statistics (so
                                   ∗
             different η’s: η1 and η1 ) the ratio of evidences may no
             longer be equal to one.’
                                        [Michael Stumpf, Jan. 28, 2011, ’Og]


      Using different summary statistics [on different models] may
      indicate the loss of information brought by each set but agreement
      does not lead to trustworthy approximations.
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




A population genetics evaluation


      Population genetics example with
             3 populations
             2 scenari
             15 individuals
             5 loci
             single mutation parameter
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




A population genetics evaluation


      Population genetics example with
             3 populations
             2 scenari
             15 individuals
             5 loci
             single mutation parameter
             24 summary statistics
             2 million ABC proposal
             importance [tree] sampling alternative
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




A population genetics evaluation


      Population genetics example with
             3 populations
             2 scenari
             15 individuals
             5 loci
             single mutation parameter
             24 summary statistics
             2 million ABC proposal
             importance [tree] sampling alternative
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




Stability of importance sampling


                             1.0




                                        1.0




                                              1.0




                                                        1.0




                                                                  1.0
                                                              q




                                   q
                             0.8




                                        0.8




                                              0.8




                                                        0.8




                                                                  0.8
                             0.6




                                        0.6




                                              0.6




                                                        0.6




                                                                  0.6
                                                    q
                             0.4




                                        0.4




                                              0.4




                                                        0.4




                                                                  0.4
                             0.2




                                        0.2




                                              0.2




                                                        0.2




                                                                  0.2
                             0.0




                                        0.0




                                              0.0




                                                        0.0




                                                                  0.0
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




Comparison with ABC
      Use of 24 summary statistics and DIY-ABC logistic correction

                                                       1.0
                                                                                                                                                                      qq
                                                                                                                                                                      qq
                                                                                                                                                                 qq qq
                                                                                                                                                                    qqq
                                                                                                                                                                 qq q q
                                                                                                                                                        qq       q q qq
                                                                                                                                                                 q q
                                                                                                                                                                     q
                                                                                                                               q            q          q         q q
                                                                                                                                                                 q
                                                                                                                                                         q qq q q
                                                                                                                                                             qq
                                                                                                                                                       q q q          qq
                                                                                                                                                qq     qq
                                                                                                                                                       q      q q q qq
                                                                                                                                                                   q q
                                                                                                                                                           q           q
                                                                                                                                                     q               q
                                                                                                                                                                    q q
                                                                                                                                            q      q
                                                                                                                                               q
                                                                                                                                                      qq
                                                                                                                                                       q      q q qq  q
                                                                                                                                                                      q
                                                       0.8




                                                                                                                                             q                      q
                                                                                                                                                 q             q     q
                                                                                                                                                   q            q
                                                                                                                                               q                 q     q
                                                                                                                                                                       q
                                                                                                                                  q                    qq q
                                                                                                                                                              q q
                                                                                                                                   q    q            q q qq q
                                                                                                                                                            q
                                                                                                                                     q    q           q qqq q q
                                                                                                                                                          qq
                                                                                                                                                        q
                             ABC direct and logistic




                                                                                                                                                       q
                                                                                                                                            q qq
                                                                                                                             q     q q       q
                                                                                                                                                                 q
                                                                                                                                                   q q
                                                       0.6




                                                                                                                                      q      q     q    qq
                                                                                                                                   q           q
                                                                                                                               qq
                                                                                                                                     q q              q q
                                                                                                                           q      q q            q
                                                                                             q                               q q
                                                                                                                              q                q
                                                                                                                   q          q      q                q q
                                                                                                 q q      q                q q
                                                                                                                             q                   q
                                                                                                                                  qq
                                                                                                               qq
                                                                                                   q      q
                                                                                     qq
                                                       0.4




                                                                                                  qq               q
                                                                                q                                      q
                                                                                                      q        q
                                                                            q                         q                q
                                                                       q                                                            q
                                                                                             q                     q
                                                                            q                         q

                                                               q                     q            q
                                                       0.2




                                                                       q                         qq
                                                                                         q         q
                                                                   q
                                                                                                  q

                                                                            q                         q
                                                                                q
                                                                       q
                                                                        q
                                                                            q
                                                       0.0




                                                               qq




                                                             0.0                    0.2                       0.4                  0.6               0.8             1.0

                                                                                                          importance sampling
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




Comparison with ABC
      Use of 15 summary statistics and DIY-ABC logistic correction

                                          6
                                          4




                                                                                                                   q                 q
                                                                                                                                 q         q
                                                                                                                           q
                                                                                                                   q
                                                                                                                   q                  qq
                                          2




                                                                                                                          q          q
                                                                                                                         q                         q
                                                                                                                           qq
                             ABC direct




                                                                                                    q      q q              qq       q
                                                                                                            qq         q
                                                                                                                           q
                                                                                                        q
                                                                                                              q                      qq
                                                                                                    qq q qq q
                                                                                                           q q
                                                                                                              q
                                                                                               q    qq q q             q
                                                                                                  q  q qq
                                                                                           q q qq
                                                                                                q    qq        q
                                                                                         q qq q q q qqq
                                                                                         q
                                                                             q     q    q q
                                                                                        q        q
                                                                                                      q
                                          0




                                                                              qq qq     q
                                                                                       q q        q
                                                                        qq     qq
                                                                    q           q qq
                                                        q       q               q
                                                                q
                                                q                             q
                                                            q
                                                    q
                                          −2
                                          −4




                                               −4               −2                 0                    2                        4             6

                                                                                  importance sampling
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




Comparison with ABC
      Use of 15 summary statistics and DIY-ABC logistic correction

                                                                                                                                     qq
                                                       6
                                                                                                                                                     q
                                                                                                                                                             q
                                                                                                                                          q
                                                                                                                                                 q
                                                                                                                                             q
                                                                                                                               q
                                                                                                                                   q q
                                                                                                                           q        q       q
                                                       4




                                                                                                                              q
                                                                                                                  qq       q qq    qq      q
                                                                                                                           q      q        q
                                                                                                                            q   q q
                                                                                                     q    q     q          qq
                                                                                                                      q                    q
                                                                                                                   q q q      q           q
                             ABC direct and logistic




                                                                                                            q   q q q   q                            q
                                                                                                           q    qq                 q    q
                                                                                                                 q      q q q q             qq
                                                                                                                    q
                                                       2




                                                                                                               q                   q       q
                                                                                                                                  q                          q
                                                                                                           qq q qq
                                                                                                                q       q q         qq
                                                                                                                                     qq    q
                                                                                                           qq            qq     q
                                                                                                                                    q
                                                                                                       q    qq q           q              qq
                                                                                                               qq qq q q
                                                                                                       qqqq         qq q q  q
                                                                                                                qqqqq
                                                                                                                 qq q           q
                                                                                                         q qq q
                                                                                                    q q q qq                q
                                                                                                     qq     q q qqq
                                                                                                                  q
                                                                                       q           qqqqq q q
                                                                                                    qq       q
                                                                                             q     qq q q    q qq
                                                       0




                                                                                        qq qqq      q q
                                                                                     qq  q qqq
                                                                                          q
                                                                                          q
                                                                                 q        q q
                                                                             q            q    q      q
                                                                     q                 q q q
                                                                             q
                                                             q                       q   q
                                                                         q               q
                                                                                      q qq
                                                                 q                       q
                                                       −2




                                                                             q            q
                                                                                 q

                                                                     q

                                                                         q
                                                       −4




                                                                             q
                                                                 q
                                                             q



                                                            −4               −2                0                    2                    4               6

                                                                                              importance sampling
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




Comparison with ABC
      Use of 15 summary statistics and DIY-ABC logistic correction

                                                                                                                                   q
                                         2




                                                                                                                  q
                                                                                                              q
                                         1




                                                                        q                         q    q
                                                                                                                                    q
                                                                       q                                q
                                                    q
                                                                   q                                      q
                                                                                          q
                             log−ratio




                                                                 q             q             q                             q
                                                                  q                                        qq
                                                            q            q            q                 qqq                q
                                               q                q                    q         q                   q         q
                                                          qq                          q     q q q                                      q
                                                 qq     qq q
                                                       qq     q        q qqqqqq q q qqqq qq
                                                                   qq q qq qq q                                     qq      q    qq
                                              qq q q                  q                                       qq q qq q q qqq
                                                                                                                       q
                                         0




                                                 q q                                         q q      qq        q                   q
                                                   q     qq    q
                                                               q    q      q      q     q            qq                             q
                                                  qq q qq qq
                                                     q           qq         qqq qq qq q     q         q q     qq     q        qqqq q
                                               q       q
                                                           q
                                                                       q
                                                                             q
                                                                                    q q          qq
                                                                                                qqq    q
                                                                                                            qq    qq q q qq
                                                                                                                                  q    q
                                              q               q     qq           q                 q     qq        q                  q
                                                q     q      q          q q              q                      qq    qqqq              q
                                                                                                             q                          q
                                                                                          q                              q    q
                                                                                                    q                          q
                                                                               q
                                         −1




                                                                                                                                   q
                                                                        q
                                                                                                              q
                                                                                            q
                                         −2




                                                                                                                                    q



                                              0               20                40                60                  80               100

                                                                                        index
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




A second population genetics experiment

             three populations, two divergent 100 gen. ago
             two scenarios [third pop. recent admixture between first two
             pop. / diverging from pop. 1 5 gen. ago]
             In scenario 1, admixture rate 0.7 from pop. 1
             100 datasets with 100 diploid individuals per population, 50
             independent microsatellite loci.
             Effective population size of 1000 and mutation rates of
             0.0005.
             6 parameters: admixture rate (U [0.1, 0.9]), three effective
             population sizes (U [200, 2000]), the time of admixture/second
             divergence (U [1, 10]) and time of first divergence (U [50, 500]).
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




Results
      IS algorithm performed with 100 coalescent trees per particle and
      50,000 particles [12 calendar days using 376 processors]
      Using ten times as many loci and seven times as many individuals
      degrades the confidence in the importance sampling approximation
      because of an increased variability in the likelihood.
                               0.9
                               0.8




                                        q
                               0.7
                               0.6




                                             q
                               0.5
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




Results
      Blurs potential divergence between ABC and genuine posterior
      probabilities because both are overwhelmingly close to one, due to
      the high information content of the data.
                                                                1.0         q qq                         q                            qq       q q qqq
                                                                                                                                                     qq
                                                                                                                                                      q
                                                                                                                                                   q qq
                                                                                                                                                     qq
                                                                                                                                                      q
                                                                                           q                                                      q qq
                                                                                                                                                     qq
                                                                                                                                                      q
                                                                                                                                                   qq
                                                                                                                                                q
                                                                                                                                           q      q
                                                                                                                                                      q

                                                                                                                                  q
                                                                0.8




                                                                                                                       q
                             Importance Sampling estimates of




                                                                                                                                      q
                               P(M=1|y) with 1,000 particles




                                                                                                                            q                     q
                                                                0.6
                                                                0.4
                                                                0.2
                                                                0.0




                                                                      0.0          0.2           0.4         0.6                0.8                   1.0
                                                                                         Importance Sampling estimates of
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




Results
      Blurs potential divergence between ABC and genuine posterior
      probabilities because both are overwhelmingly close to one, due to
      the high information content of the data.                                                1.0




                                                                                                                           q                   q qqq q qqqq
                                                                                                                                                 q q qqqqq
                                                                                                                                                   q    qqq
                                                                                                                                                         qq
                                                                                                                                                         qq
                                                                                                                                                          qq
                                                                                                                                                          qq
                                                                                                                                                          qq
                                                                                                                                                           q
                                                                                                                                                 q q qq qqq
                             Importance Sampling estimates of P(M=1|y) with 50,000 particles




                                                                                                                 q                                q       qq
                                                                                                                                                           q
                                                                                                                                               q            q
                                                                                                                                                         qq
                                                                                                                                                           q
                                                                                                                                                         q q
                                                                                                                                                           q
                                                                                                                                                           q
                                                                                               0.8




                                                                                                                                     q
                                                                                                                                                           q



                                                                                                                                                        q q
                                                                                               0.6
                                                                                               0.4
                                                                                               0.2
                                                                                               0.0




                                                                                                     0.0   0.2       0.4       0.6       0.8              1.0
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




The only safe cases



      Besides specific models like Gibbs random fields,
      using distances over the data itself escapes the discrepancy...
                                [Toni & Stumpf, 2010;Sousa et al., 2009]
ABC Methods for Bayesian Model Choice
  Generic ABC model choice




The only safe cases



      Besides specific models like Gibbs random fields,
      using distances over the data itself escapes the discrepancy...
                                [Toni & Stumpf, 2010;Sousa et al., 2009]

      ...and so does the use of more informal model fitting measures
                         [Ratmann, Andrieu, Richardson and Wiujf, 2009]

Mais conteúdo relacionado

Mais procurados

Convergence of ABC methods
Convergence of ABC methodsConvergence of ABC methods
Convergence of ABC methodsChristian Robert
 
45th SIS Meeting, Padova, Italy
45th SIS Meeting, Padova, Italy45th SIS Meeting, Padova, Italy
45th SIS Meeting, Padova, ItalyChristian Robert
 
Bayesian inference on mixtures
Bayesian inference on mixturesBayesian inference on mixtures
Bayesian inference on mixturesChristian Robert
 
ABC short course: survey chapter
ABC short course: survey chapterABC short course: survey chapter
ABC short course: survey chapterChristian Robert
 
Likelihood free computational statistics
Likelihood free computational statisticsLikelihood free computational statistics
Likelihood free computational statisticsPierre Pudlo
 
Considerate Approaches to ABC Model Selection
Considerate Approaches to ABC Model SelectionConsiderate Approaches to ABC Model Selection
Considerate Approaches to ABC Model SelectionMichael Stumpf
 
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
 
Reliable ABC model choice via random forests
Reliable ABC model choice via random forestsReliable ABC model choice via random forests
Reliable ABC model choice via random forestsChristian Robert
 
ABC short course: model choice chapter
ABC short course: model choice chapterABC short course: model choice chapter
ABC short course: model choice chapterChristian Robert
 
Approximate Bayesian model choice via random forests
Approximate Bayesian model choice via random forestsApproximate Bayesian model choice via random forests
Approximate Bayesian model choice via random forestsChristian Robert
 
from model uncertainty to ABC
from model uncertainty to ABCfrom model uncertainty to ABC
from model uncertainty to ABCChristian Robert
 
4th joint Warwick Oxford Statistics Seminar
4th joint Warwick Oxford Statistics Seminar4th joint Warwick Oxford Statistics Seminar
4th joint Warwick Oxford Statistics SeminarChristian Robert
 
Yes III: Computational methods for model choice
Yes III: Computational methods for model choiceYes III: Computational methods for model choice
Yes III: Computational methods for model choiceChristian Robert
 
ABC short course: final chapters
ABC short course: final chaptersABC short course: final chapters
ABC short course: final chaptersChristian Robert
 
Monte Carlo in Montréal 2017
Monte Carlo in Montréal 2017Monte Carlo in Montréal 2017
Monte Carlo in Montréal 2017Christian Robert
 

Mais procurados (20)

Convergence of ABC methods
Convergence of ABC methodsConvergence of ABC methods
Convergence of ABC methods
 
45th SIS Meeting, Padova, Italy
45th SIS Meeting, Padova, Italy45th SIS Meeting, Padova, Italy
45th SIS Meeting, Padova, Italy
 
Bayesian inference on mixtures
Bayesian inference on mixturesBayesian inference on mixtures
Bayesian inference on mixtures
 
ABC short course: survey chapter
ABC short course: survey chapterABC short course: survey chapter
ABC short course: survey chapter
 
Likelihood free computational statistics
Likelihood free computational statisticsLikelihood free computational statistics
Likelihood free computational statistics
 
Jsm09 talk
Jsm09 talkJsm09 talk
Jsm09 talk
 
Considerate Approaches to ABC Model Selection
Considerate Approaches to ABC Model SelectionConsiderate Approaches to ABC Model Selection
Considerate Approaches to ABC Model Selection
 
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
 
Reliable ABC model choice via random forests
Reliable ABC model choice via random forestsReliable ABC model choice via random forests
Reliable ABC model choice via random forests
 
ABC short course: model choice chapter
ABC short course: model choice chapterABC short course: model choice chapter
ABC short course: model choice chapter
 
Approximate Bayesian model choice via random forests
Approximate Bayesian model choice via random forestsApproximate Bayesian model choice via random forests
Approximate Bayesian model choice via random forests
 
Savage-Dickey paradox
Savage-Dickey paradoxSavage-Dickey paradox
Savage-Dickey paradox
 
from model uncertainty to ABC
from model uncertainty to ABCfrom model uncertainty to ABC
from model uncertainty to ABC
 
Boston talk
Boston talkBoston talk
Boston talk
 
4th joint Warwick Oxford Statistics Seminar
4th joint Warwick Oxford Statistics Seminar4th joint Warwick Oxford Statistics Seminar
4th joint Warwick Oxford Statistics Seminar
 
ABC-Gibbs
ABC-GibbsABC-Gibbs
ABC-Gibbs
 
ABC-Gibbs
ABC-GibbsABC-Gibbs
ABC-Gibbs
 
Yes III: Computational methods for model choice
Yes III: Computational methods for model choiceYes III: Computational methods for model choice
Yes III: Computational methods for model choice
 
ABC short course: final chapters
ABC short course: final chaptersABC short course: final chapters
ABC short course: final chapters
 
Monte Carlo in Montréal 2017
Monte Carlo in Montréal 2017Monte Carlo in Montréal 2017
Monte Carlo in Montréal 2017
 

Destaque

RSS Read Paper by Mark Girolami
RSS Read Paper by Mark GirolamiRSS Read Paper by Mark Girolami
RSS Read Paper by Mark GirolamiChristian Robert
 
Networks community detection using artificial bee colony swarm optimization
Networks community detection using artificial bee colony swarm optimizationNetworks community detection using artificial bee colony swarm optimization
Networks community detection using artificial bee colony swarm optimizationAboul Ella Hassanien
 
ABC short course: introduction chapters
ABC short course: introduction chaptersABC short course: introduction chapters
ABC short course: introduction chaptersChristian Robert
 
ABC Algorithm.
ABC Algorithm.ABC Algorithm.
ABC Algorithm.N Vinayak
 
Artificial bee colony (abc)
Artificial bee colony (abc)Artificial bee colony (abc)
Artificial bee colony (abc)quadmemo
 
ABC, an effective tool for selective harmonic elimination in multilevel inve...
ABC, an effective tool for selective  harmonic elimination in multilevel inve...ABC, an effective tool for selective  harmonic elimination in multilevel inve...
ABC, an effective tool for selective harmonic elimination in multilevel inve...Shridhar kulkarni
 
Artificial Bee Colony algorithm
Artificial Bee Colony algorithmArtificial Bee Colony algorithm
Artificial Bee Colony algorithmAhmed Fouad Ali
 

Destaque (9)

RSS Read Paper by Mark Girolami
RSS Read Paper by Mark GirolamiRSS Read Paper by Mark Girolami
RSS Read Paper by Mark Girolami
 
Networks community detection using artificial bee colony swarm optimization
Networks community detection using artificial bee colony swarm optimizationNetworks community detection using artificial bee colony swarm optimization
Networks community detection using artificial bee colony swarm optimization
 
ABC workshop: 17w5025
ABC workshop: 17w5025ABC workshop: 17w5025
ABC workshop: 17w5025
 
ABC short course: introduction chapters
ABC short course: introduction chaptersABC short course: introduction chapters
ABC short course: introduction chapters
 
ABC Algorithm.
ABC Algorithm.ABC Algorithm.
ABC Algorithm.
 
Artificial bee colony (abc)
Artificial bee colony (abc)Artificial bee colony (abc)
Artificial bee colony (abc)
 
ABC, an effective tool for selective harmonic elimination in multilevel inve...
ABC, an effective tool for selective  harmonic elimination in multilevel inve...ABC, an effective tool for selective  harmonic elimination in multilevel inve...
ABC, an effective tool for selective harmonic elimination in multilevel inve...
 
Abc analysis
Abc analysisAbc analysis
Abc analysis
 
Artificial Bee Colony algorithm
Artificial Bee Colony algorithmArtificial Bee Colony algorithm
Artificial Bee Colony algorithm
 

Semelhante a ABC in London, May 5, 2011

Approximating Bayes Factors
Approximating Bayes FactorsApproximating Bayes Factors
Approximating Bayes FactorsChristian Robert
 
(Approximate) Bayesian computation as a new empirical Bayes (something)?
(Approximate) Bayesian computation as a new empirical Bayes (something)?(Approximate) Bayesian computation as a new empirical Bayes (something)?
(Approximate) Bayesian computation as a new empirical Bayes (something)?Christian Robert
 
Is ABC a new empirical Bayes approach?
Is ABC a new empirical Bayes approach?Is ABC a new empirical Bayes approach?
Is ABC a new empirical Bayes approach?Christian Robert
 
ABC and empirical likelihood
ABC and empirical likelihoodABC and empirical likelihood
ABC and empirical likelihoodChristian Robert
 
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
 
Semi-automatic ABC: a discussion
Semi-automatic ABC: a discussionSemi-automatic ABC: a discussion
Semi-automatic ABC: a discussionChristian Robert
 
NBBC15, Reyjavik, June 08, 2015
NBBC15, Reyjavik, June 08, 2015NBBC15, Reyjavik, June 08, 2015
NBBC15, Reyjavik, June 08, 2015Christian Robert
 
ABC and empirical likelihood
ABC and empirical likelihoodABC and empirical likelihood
ABC and empirical likelihoodChristian Robert
 

Semelhante a ABC in London, May 5, 2011 (17)

Intro to ABC
Intro to ABCIntro to ABC
Intro to ABC
 
Desy
DesyDesy
Desy
 
Approximating Bayes Factors
Approximating Bayes FactorsApproximating Bayes Factors
Approximating Bayes Factors
 
(Approximate) Bayesian computation as a new empirical Bayes (something)?
(Approximate) Bayesian computation as a new empirical Bayes (something)?(Approximate) Bayesian computation as a new empirical Bayes (something)?
(Approximate) Bayesian computation as a new empirical Bayes (something)?
 
MaxEnt 2009 talk
MaxEnt 2009 talkMaxEnt 2009 talk
MaxEnt 2009 talk
 
Is ABC a new empirical Bayes approach?
Is ABC a new empirical Bayes approach?Is ABC a new empirical Bayes approach?
Is ABC a new empirical Bayes approach?
 
ABC in Varanasi
ABC in VaranasiABC in Varanasi
ABC in Varanasi
 
Maths Assignment Help
Maths Assignment HelpMaths Assignment Help
Maths Assignment Help
 
ABC and empirical likelihood
ABC and empirical likelihoodABC and empirical likelihood
ABC and empirical likelihood
 
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
 
MUMS: Bayesian, Fiducial, and Frequentist Conference - Inference on Treatment...
MUMS: Bayesian, Fiducial, and Frequentist Conference - Inference on Treatment...MUMS: Bayesian, Fiducial, and Frequentist Conference - Inference on Treatment...
MUMS: Bayesian, Fiducial, and Frequentist Conference - Inference on Treatment...
 
Semi-automatic ABC: a discussion
Semi-automatic ABC: a discussionSemi-automatic ABC: a discussion
Semi-automatic ABC: a discussion
 
NBBC15, Reyjavik, June 08, 2015
NBBC15, Reyjavik, June 08, 2015NBBC15, Reyjavik, June 08, 2015
NBBC15, Reyjavik, June 08, 2015
 
Intractable likelihoods
Intractable likelihoodsIntractable likelihoods
Intractable likelihoods
 
ABC and empirical likelihood
ABC and empirical likelihoodABC and empirical likelihood
ABC and empirical likelihood
 
ABC-Gibbs
ABC-GibbsABC-Gibbs
ABC-Gibbs
 
MUMS: Bayesian, Fiducial, and Frequentist Conference - Including Factors in B...
MUMS: Bayesian, Fiducial, and Frequentist Conference - Including Factors in B...MUMS: Bayesian, Fiducial, and Frequentist Conference - Including Factors in B...
MUMS: Bayesian, Fiducial, and Frequentist Conference - Including Factors in B...
 

Mais de Christian 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
 
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
 
How many components in a mixture?
How many components in a mixture?How many components in a mixture?
How many components in a mixture?Christian Robert
 
Testing for mixtures at BNP 13
Testing for mixtures at BNP 13Testing for mixtures at BNP 13
Testing for mixtures at BNP 13Christian Robert
 
Inferring the number of components: dream or reality?
Inferring the number of components: dream or reality?Inferring the number of components: dream or reality?
Inferring the number of components: dream or reality?Christian Robert
 
Testing for mixtures by seeking components
Testing for mixtures by seeking componentsTesting for mixtures by seeking components
Testing for mixtures by seeking componentsChristian Robert
 
discussion on Bayesian restricted likelihood
discussion on Bayesian restricted likelihooddiscussion on Bayesian restricted likelihood
discussion on Bayesian restricted likelihoodChristian Robert
 
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
 
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
 
Laplace's Demon: seminar #1
Laplace's Demon: seminar #1Laplace's Demon: seminar #1
Laplace's Demon: seminar #1Christian Robert
 
Likelihood-free Design: a discussion
Likelihood-free Design: a discussionLikelihood-free Design: a discussion
Likelihood-free Design: a discussionChristian Robert
 
CISEA 2019: ABC consistency and convergence
CISEA 2019: ABC consistency and convergenceCISEA 2019: ABC consistency and convergence
CISEA 2019: ABC consistency and convergenceChristian Robert
 
a discussion of Chib, Shin, and Simoni (2017-8) Bayesian moment models
a discussion of Chib, Shin, and Simoni (2017-8) Bayesian moment modelsa discussion of Chib, Shin, and Simoni (2017-8) Bayesian moment models
a discussion of Chib, Shin, and Simoni (2017-8) Bayesian moment modelsChristian Robert
 
ABC based on Wasserstein distances
ABC based on Wasserstein distancesABC based on Wasserstein distances
ABC based on Wasserstein distancesChristian Robert
 

Mais de Christian Robert (20)

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
 
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
 
discussion of ICML23.pdf
discussion of ICML23.pdfdiscussion of ICML23.pdf
discussion of ICML23.pdf
 
How many components in a mixture?
How many components in a mixture?How many components in a mixture?
How many components in a mixture?
 
restore.pdf
restore.pdfrestore.pdf
restore.pdf
 
Testing for mixtures at BNP 13
Testing for mixtures at BNP 13Testing for mixtures at BNP 13
Testing for mixtures at BNP 13
 
Inferring the number of components: dream or reality?
Inferring the number of components: dream or reality?Inferring the number of components: dream or reality?
Inferring the number of components: dream or reality?
 
CDT 22 slides.pdf
CDT 22 slides.pdfCDT 22 slides.pdf
CDT 22 slides.pdf
 
Testing for mixtures by seeking components
Testing for mixtures by seeking componentsTesting for mixtures by seeking components
Testing for mixtures by seeking components
 
discussion on Bayesian restricted likelihood
discussion on Bayesian restricted likelihooddiscussion on Bayesian restricted likelihood
discussion on Bayesian restricted likelihood
 
NCE, GANs & VAEs (and maybe BAC)
NCE, GANs & VAEs (and maybe BAC)NCE, GANs & VAEs (and maybe BAC)
NCE, GANs & VAEs (and maybe BAC)
 
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
 
eugenics and statistics
eugenics and statisticseugenics and statistics
eugenics and statistics
 
Laplace's Demon: seminar #1
Laplace's Demon: seminar #1Laplace's Demon: seminar #1
Laplace's Demon: seminar #1
 
asymptotics of ABC
asymptotics of ABCasymptotics of ABC
asymptotics of ABC
 
Likelihood-free Design: a discussion
Likelihood-free Design: a discussionLikelihood-free Design: a discussion
Likelihood-free Design: a discussion
 
the ABC of ABC
the ABC of ABCthe ABC of ABC
the ABC of ABC
 
CISEA 2019: ABC consistency and convergence
CISEA 2019: ABC consistency and convergenceCISEA 2019: ABC consistency and convergence
CISEA 2019: ABC consistency and convergence
 
a discussion of Chib, Shin, and Simoni (2017-8) Bayesian moment models
a discussion of Chib, Shin, and Simoni (2017-8) Bayesian moment modelsa discussion of Chib, Shin, and Simoni (2017-8) Bayesian moment models
a discussion of Chib, Shin, and Simoni (2017-8) Bayesian moment models
 
ABC based on Wasserstein distances
ABC based on Wasserstein distancesABC based on Wasserstein distances
ABC based on Wasserstein distances
 

Último

How to setup Pycharm environment for Odoo 17.pptx
How to setup Pycharm environment for Odoo 17.pptxHow to setup Pycharm environment for Odoo 17.pptx
How to setup Pycharm environment for Odoo 17.pptxCeline George
 
Interdisciplinary_Insights_Data_Collection_Methods.pptx
Interdisciplinary_Insights_Data_Collection_Methods.pptxInterdisciplinary_Insights_Data_Collection_Methods.pptx
Interdisciplinary_Insights_Data_Collection_Methods.pptxPooja Bhuva
 
General Principles of Intellectual Property: Concepts of Intellectual Proper...
General Principles of Intellectual Property: Concepts of Intellectual  Proper...General Principles of Intellectual Property: Concepts of Intellectual  Proper...
General Principles of Intellectual Property: Concepts of Intellectual Proper...Poonam Aher Patil
 
How to Add New Custom Addons Path in Odoo 17
How to Add New Custom Addons Path in Odoo 17How to Add New Custom Addons Path in Odoo 17
How to Add New Custom Addons Path in Odoo 17Celine George
 
Graduate Outcomes Presentation Slides - English
Graduate Outcomes Presentation Slides - EnglishGraduate Outcomes Presentation Slides - English
Graduate Outcomes Presentation Slides - Englishneillewis46
 
Jamworks pilot and AI at Jisc (20/03/2024)
Jamworks pilot and AI at Jisc (20/03/2024)Jamworks pilot and AI at Jisc (20/03/2024)
Jamworks pilot and AI at Jisc (20/03/2024)Jisc
 
Basic Civil Engineering first year Notes- Chapter 4 Building.pptx
Basic Civil Engineering first year Notes- Chapter 4 Building.pptxBasic Civil Engineering first year Notes- Chapter 4 Building.pptx
Basic Civil Engineering first year Notes- Chapter 4 Building.pptxDenish Jangid
 
Single or Multiple melodic lines structure
Single or Multiple melodic lines structureSingle or Multiple melodic lines structure
Single or Multiple melodic lines structuredhanjurrannsibayan2
 
Understanding Accommodations and Modifications
Understanding  Accommodations and ModificationsUnderstanding  Accommodations and Modifications
Understanding Accommodations and ModificationsMJDuyan
 
UGC NET Paper 1 Mathematical Reasoning & Aptitude.pdf
UGC NET Paper 1 Mathematical Reasoning & Aptitude.pdfUGC NET Paper 1 Mathematical Reasoning & Aptitude.pdf
UGC NET Paper 1 Mathematical Reasoning & Aptitude.pdfNirmal Dwivedi
 
Key note speaker Neum_Admir Softic_ENG.pdf
Key note speaker Neum_Admir Softic_ENG.pdfKey note speaker Neum_Admir Softic_ENG.pdf
Key note speaker Neum_Admir Softic_ENG.pdfAdmir Softic
 
Fostering Friendships - Enhancing Social Bonds in the Classroom
Fostering Friendships - Enhancing Social Bonds  in the ClassroomFostering Friendships - Enhancing Social Bonds  in the Classroom
Fostering Friendships - Enhancing Social Bonds in the ClassroomPooky Knightsmith
 
Exploring_the_Narrative_Style_of_Amitav_Ghoshs_Gun_Island.pptx
Exploring_the_Narrative_Style_of_Amitav_Ghoshs_Gun_Island.pptxExploring_the_Narrative_Style_of_Amitav_Ghoshs_Gun_Island.pptx
Exploring_the_Narrative_Style_of_Amitav_Ghoshs_Gun_Island.pptxPooja Bhuva
 
Accessible Digital Futures project (20/03/2024)
Accessible Digital Futures project (20/03/2024)Accessible Digital Futures project (20/03/2024)
Accessible Digital Futures project (20/03/2024)Jisc
 
Towards a code of practice for AI in AT.pptx
Towards a code of practice for AI in AT.pptxTowards a code of practice for AI in AT.pptx
Towards a code of practice for AI in AT.pptxJisc
 
Micro-Scholarship, What it is, How can it help me.pdf
Micro-Scholarship, What it is, How can it help me.pdfMicro-Scholarship, What it is, How can it help me.pdf
Micro-Scholarship, What it is, How can it help me.pdfPoh-Sun Goh
 
Plant propagation: Sexual and Asexual propapagation.pptx
Plant propagation: Sexual and Asexual propapagation.pptxPlant propagation: Sexual and Asexual propapagation.pptx
Plant propagation: Sexual and Asexual propapagation.pptxUmeshTimilsina1
 
How to Give a Domain for a Field in Odoo 17
How to Give a Domain for a Field in Odoo 17How to Give a Domain for a Field in Odoo 17
How to Give a Domain for a Field in Odoo 17Celine George
 
This PowerPoint helps students to consider the concept of infinity.
This PowerPoint helps students to consider the concept of infinity.This PowerPoint helps students to consider the concept of infinity.
This PowerPoint helps students to consider the concept of infinity.christianmathematics
 

Último (20)

Mehran University Newsletter Vol-X, Issue-I, 2024
Mehran University Newsletter Vol-X, Issue-I, 2024Mehran University Newsletter Vol-X, Issue-I, 2024
Mehran University Newsletter Vol-X, Issue-I, 2024
 
How to setup Pycharm environment for Odoo 17.pptx
How to setup Pycharm environment for Odoo 17.pptxHow to setup Pycharm environment for Odoo 17.pptx
How to setup Pycharm environment for Odoo 17.pptx
 
Interdisciplinary_Insights_Data_Collection_Methods.pptx
Interdisciplinary_Insights_Data_Collection_Methods.pptxInterdisciplinary_Insights_Data_Collection_Methods.pptx
Interdisciplinary_Insights_Data_Collection_Methods.pptx
 
General Principles of Intellectual Property: Concepts of Intellectual Proper...
General Principles of Intellectual Property: Concepts of Intellectual  Proper...General Principles of Intellectual Property: Concepts of Intellectual  Proper...
General Principles of Intellectual Property: Concepts of Intellectual Proper...
 
How to Add New Custom Addons Path in Odoo 17
How to Add New Custom Addons Path in Odoo 17How to Add New Custom Addons Path in Odoo 17
How to Add New Custom Addons Path in Odoo 17
 
Graduate Outcomes Presentation Slides - English
Graduate Outcomes Presentation Slides - EnglishGraduate Outcomes Presentation Slides - English
Graduate Outcomes Presentation Slides - English
 
Jamworks pilot and AI at Jisc (20/03/2024)
Jamworks pilot and AI at Jisc (20/03/2024)Jamworks pilot and AI at Jisc (20/03/2024)
Jamworks pilot and AI at Jisc (20/03/2024)
 
Basic Civil Engineering first year Notes- Chapter 4 Building.pptx
Basic Civil Engineering first year Notes- Chapter 4 Building.pptxBasic Civil Engineering first year Notes- Chapter 4 Building.pptx
Basic Civil Engineering first year Notes- Chapter 4 Building.pptx
 
Single or Multiple melodic lines structure
Single or Multiple melodic lines structureSingle or Multiple melodic lines structure
Single or Multiple melodic lines structure
 
Understanding Accommodations and Modifications
Understanding  Accommodations and ModificationsUnderstanding  Accommodations and Modifications
Understanding Accommodations and Modifications
 
UGC NET Paper 1 Mathematical Reasoning & Aptitude.pdf
UGC NET Paper 1 Mathematical Reasoning & Aptitude.pdfUGC NET Paper 1 Mathematical Reasoning & Aptitude.pdf
UGC NET Paper 1 Mathematical Reasoning & Aptitude.pdf
 
Key note speaker Neum_Admir Softic_ENG.pdf
Key note speaker Neum_Admir Softic_ENG.pdfKey note speaker Neum_Admir Softic_ENG.pdf
Key note speaker Neum_Admir Softic_ENG.pdf
 
Fostering Friendships - Enhancing Social Bonds in the Classroom
Fostering Friendships - Enhancing Social Bonds  in the ClassroomFostering Friendships - Enhancing Social Bonds  in the Classroom
Fostering Friendships - Enhancing Social Bonds in the Classroom
 
Exploring_the_Narrative_Style_of_Amitav_Ghoshs_Gun_Island.pptx
Exploring_the_Narrative_Style_of_Amitav_Ghoshs_Gun_Island.pptxExploring_the_Narrative_Style_of_Amitav_Ghoshs_Gun_Island.pptx
Exploring_the_Narrative_Style_of_Amitav_Ghoshs_Gun_Island.pptx
 
Accessible Digital Futures project (20/03/2024)
Accessible Digital Futures project (20/03/2024)Accessible Digital Futures project (20/03/2024)
Accessible Digital Futures project (20/03/2024)
 
Towards a code of practice for AI in AT.pptx
Towards a code of practice for AI in AT.pptxTowards a code of practice for AI in AT.pptx
Towards a code of practice for AI in AT.pptx
 
Micro-Scholarship, What it is, How can it help me.pdf
Micro-Scholarship, What it is, How can it help me.pdfMicro-Scholarship, What it is, How can it help me.pdf
Micro-Scholarship, What it is, How can it help me.pdf
 
Plant propagation: Sexual and Asexual propapagation.pptx
Plant propagation: Sexual and Asexual propapagation.pptxPlant propagation: Sexual and Asexual propapagation.pptx
Plant propagation: Sexual and Asexual propapagation.pptx
 
How to Give a Domain for a Field in Odoo 17
How to Give a Domain for a Field in Odoo 17How to Give a Domain for a Field in Odoo 17
How to Give a Domain for a Field in Odoo 17
 
This PowerPoint helps students to consider the concept of infinity.
This PowerPoint helps students to consider the concept of infinity.This PowerPoint helps students to consider the concept of infinity.
This PowerPoint helps students to consider the concept of infinity.
 

ABC in London, May 5, 2011

  • 1. ABC Methods for Bayesian Model Choice ABC Methods for Bayesian Model Choice Christian P. Robert Universit´ Paris-Dauphine, IuF, & CREST e http://www.ceremade.dauphine.fr/~xian May 4, 2011
  • 2. ABC Methods for Bayesian Model Choice ABC for model choice Model choice Gibbs random fields Generic ABC model choice
  • 3. ABC Methods for Bayesian Model Choice Model choice Bayesian model choice Several models M1 , M2 , . . . are considered simultaneously for a dataset y and the model index M is part of the inference. Use of a prior distribution. π(M = m), plus a prior distribution on the parameter conditional on the value m of the model index, πm (θ m ) Goal is to derive the posterior distribution of M , challenging computational target when models are complex.
  • 4. ABC Methods for Bayesian Model Choice Model choice Generic ABC for model choice Algorithm 1 Likelihood-free model choice sampler (ABC-MC) for t = 1 to T do repeat Generate m from the prior π(M = m) Generate θ m from the prior πm (θ m ) Generate z from the model fm (z|θ m ) until ρ{η(z), η(y)} < Set m(t) = m and θ (t) = θ m end for [Toni, Welch, Strelkowa, Ipsen & Stumpf, 2009]
  • 5. ABC Methods for Bayesian Model Choice Model choice ABC estimates Posterior probability π(M = m|y) approximated by the frequency of acceptances from model m T 1 Im(t) =m . T t=1 Early issues with implementation: should tolerances be the same for all models? should summary statistics vary across models? (incl. their dimension)? should the distance measure ρ vary as well?
  • 6. ABC Methods for Bayesian Model Choice Model choice ABC estimates Posterior probability π(M = m|y) approximated by the frequency of acceptances from model m T 1 Im(t) =m . T t=1 Early issues with implementation: then needs to become part of the model this is incompatible with Bayesian model choice same difficulty Extension to a weighted polychotomous logistic regression estimate of π(M = m|y), with non-parametric kernel weights [Cornuet et al., DIYABC, 2009]
  • 7. ABC Methods for Bayesian Model Choice Model choice The Great ABC controversy On-going controvery in phylogeographic genetics about the validity of using ABC for testing Against: Templeton, 2008, 2009, 2010a, 2010b, 2010c argues that nested hypotheses cannot have higher probabilities than nesting hypotheses (!)
  • 8. ABC Methods for Bayesian Model Choice Model choice The Great ABC controversy On-going controvery in phylogeographic genetics about the validity of using ABC for testing Replies: Fagundes et al., 2008, Against: Templeton, 2008, Beaumont et al., 2010, Berger et 2009, 2010a, 2010b, 2010c al., 2010, Csill`ry et al., 2010 e argues that nested hypotheses point out that the criticisms are cannot have higher probabilities addressed at [Bayesian] than nesting hypotheses (!) model-based inference and have nothing to do with ABC...
  • 9. ABC Methods for Bayesian Model Choice Gibbs random fields Gibbs random fields Gibbs distribution The rv y = (y1 , . . . , yn ) is a Gibbs random field associated with the graph G if 1 f (y) = exp − Vc (yc ) , Z c∈C where Z is the normalising constant, C is the set of cliques of G and Vc is any function also called potential U (y) = c∈C Vc (yc ) is the energy function
  • 10. ABC Methods for Bayesian Model Choice Gibbs random fields Gibbs random fields Gibbs distribution The rv y = (y1 , . . . , yn ) is a Gibbs random field associated with the graph G if 1 f (y) = exp − Vc (yc ) , Z c∈C where Z is the normalising constant, C is the set of cliques of G and Vc is any function also called potential U (y) = c∈C Vc (yc ) is the energy function c Z is usually unavailable in closed form
  • 11. ABC Methods for Bayesian Model Choice Gibbs random fields Potts model Potts model Vc (y) is of the form Vc (y) = θS(y) = θ δyl =yi l∼i where l∼i denotes a neighbourhood structure
  • 12. ABC Methods for Bayesian Model Choice Gibbs random fields Potts model Potts model Vc (y) is of the form Vc (y) = θS(y) = θ δyl =yi l∼i where l∼i denotes a neighbourhood structure In most realistic settings, summation Zθ = exp{θ T S(x)} x∈X involves too many terms to be manageable and numerical approximations cannot always be trusted
  • 13. ABC Methods for Bayesian Model Choice Gibbs random fields Bayesian Model Choice Comparing a model with energy S0 taking values in Rp0 versus a model with energy S1 taking values in Rp1 can be done through the Bayes factor corresponding to the priors π0 and π1 on each parameter space exp{θ T S0 (x)}/Zθ 0 ,0 π0 (dθ 0 ) 0 Bm0 /m1 (x) = exp{θ T S1 (x)}/Zθ 1 ,1 π1 (dθ 1 ) 1
  • 14. ABC Methods for Bayesian Model Choice Gibbs random fields Neighbourhood relations Choice to be made between M neighbourhood relations m i∼i (0 ≤ m ≤ M − 1) with Sm (x) = I{xi =xi } m i∼i driven by the posterior probabilities of the models.
  • 15. ABC Methods for Bayesian Model Choice Gibbs random fields Model index Computational target: P(M = m|x) ∝ fm (x|θm )πm (θm ) dθm π(M = m) , Θm
  • 16. ABC Methods for Bayesian Model Choice Gibbs random fields Model index Computational target: P(M = m|x) ∝ fm (x|θm )πm (θm ) dθm π(M = m) , Θm If S(x) sufficient statistic for the joint parameters (M, θ0 , . . . , θM −1 ), P(M = m|x) = P(M = m|S(x)) .
  • 17. ABC Methods for Bayesian Model Choice Gibbs random fields Sufficient statistics in Gibbs random fields
  • 18. ABC Methods for Bayesian Model Choice Gibbs random fields Sufficient statistics in Gibbs random fields Each model m has its own sufficient statistic Sm (·) and S(·) = (S0 (·), . . . , SM −1 (·)) is also (model-)sufficient.
  • 19. ABC Methods for Bayesian Model Choice Gibbs random fields Sufficient statistics in Gibbs random fields Each model m has its own sufficient statistic Sm (·) and S(·) = (S0 (·), . . . , SM −1 (·)) is also (model-)sufficient. For Gibbs random fields, 1 2 x|M = m ∼ fm (x|θm ) = fm (x|S(x))fm (S(x)|θm ) 1 = f 2 (S(x)|θm ) n(S(x)) m where n(S(x)) = {˜ ∈ X : S(˜ ) = S(x)} x x c S(x) is therefore also sufficient for the joint parameters
  • 20. ABC Methods for Bayesian Model Choice Gibbs random fields ABC model choice Algorithm ABC-MC Generate m∗ from the prior π(M = m). ∗ Generate θm∗ from the prior πm∗ (·). Generate x∗ from the model fm∗ (·|θm∗ ). ∗ Compute the distance ρ(S(x0 ), S(x∗ )). Accept (θm∗ , m∗ ) if ρ(S(x0 ), S(x∗ )) < . ∗ Note When = 0 the algorithm is exact
  • 21. ABC Methods for Bayesian Model Choice Gibbs random fields Toy example iid Bernoulli model versus two-state first-order Markov chain, i.e. n f0 (x|θ0 ) = exp θ0 I{xi =1} {1 + exp(θ0 )}n , i=1 versus n 1 f1 (x|θ1 ) = exp θ1 I{xi =xi−1 } {1 + exp(θ1 )}n−1 , 2 i=2 with priors θ0 ∼ U(−5, 5) and θ1 ∼ U(0, 6) (inspired by “phase transition” boundaries).
  • 22. ABC Methods for Bayesian Model Choice Gibbs random fields Toy example (2) ^ BF01 −5 0 5 −40 −20 BF01 0 10 ^ BF01 −10 −5 0 5 10 −40 −20 BF01 0 10 (left) Comparison of the true BF m0 /m1 (x0 ) with BF m0 /m1 (x0 ) (in logs) over 2, 000 simulations and 4.106 proposals from the prior. (right) Same when using tolerance corresponding to the 1% quantile on the distances.
  • 23. ABC Methods for Bayesian Model Choice Generic ABC model choice Back to sufficiency ‘Sufficient statistics for individual models are unlikely to be very informative for the model probability. This is already well known and understood by the ABC-user community.’ [Scott Sisson, Jan. 31, 2011, ’Og]
  • 24. ABC Methods for Bayesian Model Choice Generic ABC model choice Back to sufficiency ‘Sufficient statistics for individual models are unlikely to be very informative for the model probability. This is already well known and understood by the ABC-user community.’ [Scott Sisson, Jan. 31, 2011, ’Og] If η1 (x) sufficient statistic for model m = 1 and parameter θ1 and η2 (x) sufficient statistic for model m = 2 and parameter θ2 , (η1 (x), η2 (x)) is not always sufficient for (m, θm )
  • 25. ABC Methods for Bayesian Model Choice Generic ABC model choice Back to sufficiency ‘Sufficient statistics for individual models are unlikely to be very informative for the model probability. This is already well known and understood by the ABC-user community.’ [Scott Sisson, Jan. 31, 2011, ’Og] If η1 (x) sufficient statistic for model m = 1 and parameter θ1 and η2 (x) sufficient statistic for model m = 2 and parameter θ2 , (η1 (x), η2 (x)) is not always sufficient for (m, θm ) c Potential loss of information at the testing level
  • 26. ABC Methods for Bayesian Model Choice Generic ABC model choice Limiting behaviour of B12 (T → ∞) ABC approximation T t=1 Imt =1 Iρ{η(zt ),η(y)}≤ B12 (y) = T , t=1 Imt =2 Iρ{η(zt ),η(y)}≤ where the (mt , z t )’s are simulated from the (joint) prior
  • 27. ABC Methods for Bayesian Model Choice Generic ABC model choice Limiting behaviour of B12 (T → ∞) ABC approximation T t=1 Imt =1 Iρ{η(zt ),η(y)}≤ B12 (y) = T , t=1 Imt =2 Iρ{η(zt ),η(y)}≤ where the (mt , z t )’s are simulated from the (joint) prior As T go to infinity, limit Iρ{η(z),η(y)}≤ π1 (θ 1 )f1 (z|θ 1 ) dz dθ 1 B12 (y) = Iρ{η(z),η(y)}≤ π2 (θ 2 )f2 (z|θ 2 ) dz dθ 2 η Iρ{η,η(y)}≤ π1 (θ 1 )f1 (η|θ 1 ) dη dθ 1 = η , Iρ{η,η(y)}≤ π2 (θ 2 )f2 (η|θ 2 ) dη dθ 2 η η where f1 (η|θ 1 ) and f2 (η|θ 2 ) distributions of η(z)
  • 28. ABC Methods for Bayesian Model Choice Generic ABC model choice Limiting behaviour of B12 ( → 0) When goes to zero, η η π1 (θ 1 )f1 (η(y)|θ 1 ) dθ 1 B12 (y) = η , π2 (θ 2 )f2 (η(y)|θ 2 ) dθ 2
  • 29. ABC Methods for Bayesian Model Choice Generic ABC model choice Limiting behaviour of B12 ( → 0) When goes to zero, η η π1 (θ 1 )f1 (η(y)|θ 1 ) dθ 1 B12 (y) = η , π2 (θ 2 )f2 (η(y)|θ 2 ) dθ 2 Bayes factor based on the sole observation of η(y)
  • 30. ABC Methods for Bayesian Model Choice Generic ABC model choice Limiting behaviour of B12 (under sufficiency) If η(y) sufficient statistic for both models, fi (y|θ i ) = gi (y)fiη (η(y)|θ i ) Thus η Θ1 π(θ 1 )g1 (y)f1 (η(y)|θ 1 ) dθ 1 B12 (y) = η Θ2 π(θ 2 )g2 (y)f2 (η(y)|θ 2 ) dθ 2 η g1 (y) π1 (θ 1 )f1 (η(y)|θ 1 ) dθ 1 = η g2 (y) π2 (θ 2 )f2 (η(y)|θ 2 ) dθ 2 g1 (y) η = B (y) . g2 (y) 12 [Didelot, Everitt, Johansen & Lawson, 2011]
  • 31. ABC Methods for Bayesian Model Choice Generic ABC model choice Limiting behaviour of B12 (under sufficiency) If η(y) sufficient statistic for both models, fi (y|θ i ) = gi (y)fiη (η(y)|θ i ) Thus η Θ1 π(θ 1 )g1 (y)f1 (η(y)|θ 1 ) dθ 1 B12 (y) = η Θ2 π(θ 2 )g2 (y)f2 (η(y)|θ 2 ) dθ 2 η g1 (y) π1 (θ 1 )f1 (η(y)|θ 1 ) dθ 1 = η g2 (y) π2 (θ 2 )f2 (η(y)|θ 2 ) dθ 2 g1 (y) η = B (y) . g2 (y) 12 [Didelot, Everitt, Johansen & Lawson, 2011] No discrepancy only when cross-model sufficiency
  • 32. ABC Methods for Bayesian Model Choice Generic ABC model choice Poisson/geometric example Sample x = (x1 , . . . , xn ) from either a Poisson P(λ) or from a geometric G(p) Sum n S= yi = η(x) i=1 sufficient statistic for either model but not simultaneously Discrepancy ratio g1 (x) S!n−S / i yi ! = g2 (x) 1 n+S−1 S
  • 33. ABC Methods for Bayesian Model Choice Generic ABC model choice Poisson/geometric discrepancy η Range of B12 (x) versus B12 (x) B12 (x): The values produced have nothing in common.
  • 34. ABC Methods for Bayesian Model Choice Generic ABC model choice Formal recovery Creating an encompassing exponential family T T f (x|θ1 , θ2 , α1 , α2 ) ∝ exp{θ1 η1 (x) + θ1 η1 (x) + α1 t1 (x) + α2 t2 (x)} leads to a sufficient statistic (η1 (x), η2 (x), t1 (x), t2 (x)) [Didelot, Everitt, Johansen & Lawson, 2011]
  • 35. ABC Methods for Bayesian Model Choice Generic ABC model choice Formal recovery Creating an encompassing exponential family T T f (x|θ1 , θ2 , α1 , α2 ) ∝ exp{θ1 η1 (x) + θ1 η1 (x) + α1 t1 (x) + α2 t2 (x)} leads to a sufficient statistic (η1 (x), η2 (x), t1 (x), t2 (x)) [Didelot, Everitt, Johansen & Lawson, 2011] In the Poisson/geometric case, if i xi ! is added to S, no discrepancy
  • 36. ABC Methods for Bayesian Model Choice Generic ABC model choice Formal recovery Creating an encompassing exponential family T T f (x|θ1 , θ2 , α1 , α2 ) ∝ exp{θ1 η1 (x) + θ1 η1 (x) + α1 t1 (x) + α2 t2 (x)} leads to a sufficient statistic (η1 (x), η2 (x), t1 (x), t2 (x)) [Didelot, Everitt, Johansen & Lawson, 2011] Only applies in genuine sufficiency settings... c Inability to evaluate loss brought by summary statistics
  • 37. ABC Methods for Bayesian Model Choice Generic ABC model choice The Pitman–Koopman lemma Lemma A necessary and sufficient condition for the existence of a sufficient statistic of a fixed dimension whatever the sample size is that the sampling distribution belongs to an exponential family. [Pitman, 1933; Koopman, 1933]
  • 38. ABC Methods for Bayesian Model Choice Generic ABC model choice The Pitman–Koopman lemma Lemma A necessary and sufficient condition for the existence of a sufficient statistic of a fixed dimension whatever the sample size is that the sampling distribution belongs to an exponential family. [Pitman, 1933; Koopman, 1933] Provision of fixed support (consider U(0, θ) counterexample)
  • 39. ABC Methods for Bayesian Model Choice Generic ABC model choice Meaning of the ABC-Bayes factor ‘This is also why focus on model discrimination typically (...) proceeds by (...) accepting that the Bayes Factor that one obtains is only derived from the summary statistics and may in no way correspond to that of the full model.’ [Scott Sisson, Jan. 31, 2011, ’Og]
  • 40. ABC Methods for Bayesian Model Choice Generic ABC model choice Meaning of the ABC-Bayes factor ‘This is also why focus on model discrimination typically (...) proceeds by (...) accepting that the Bayes Factor that one obtains is only derived from the summary statistics and may in no way correspond to that of the full model.’ [Scott Sisson, Jan. 31, 2011, ’Og] In the Poisson/geometric case, if E[yi ] = θ0 > 0, η (θ0 + 1)2 −θ0 lim B12 (y) = e n→∞ θ0
  • 41. ABC Methods for Bayesian Model Choice Generic ABC model choice MA example Consider the MA(q) model q xt = t+ ϑi t−i i=1 Simple prior: uniform prior over the identifiability zone, e.g. triangle for MA(2)
  • 42. ABC Methods for Bayesian Model Choice Generic ABC model choice MA example (2) ABC algorithm thus made of 1. picking a new value (ϑ1 , ϑ2 ) in the triangle 2. generating an iid sequence ( t )−q<t≤T 3. producing a simulated series (xt )1≤t≤T
  • 43. ABC Methods for Bayesian Model Choice Generic ABC model choice MA example (2) ABC algorithm thus made of 1. picking a new value (ϑ1 , ϑ2 ) in the triangle 2. generating an iid sequence ( t )−q<t≤T 3. producing a simulated series (xt )1≤t≤T Distance: basic distance between the series T ρ((xt )1≤t≤T , (xt )1≤t≤T ) = (xt − xt )2 t=1 or between summary statistics like the first q autocorrelations T τj = xt xt−j t=j+1
  • 44. ABC Methods for Bayesian Model Choice Generic ABC model choice Comparison of distance impact Evaluation of the tolerance on the ABC sample against both distances ( = 100%, 10%, 1%, 0.1%) for an MA(2) model
  • 45. ABC Methods for Bayesian Model Choice Generic ABC model choice Comparison of distance impact 4 1.5 3 1.0 2 0.5 1 0.0 0 0.0 0.2 0.4 0.6 0.8 −2.0 −1.0 0.0 0.5 1.0 1.5 θ1 θ2 Evaluation of the tolerance on the ABC sample against both distances ( = 100%, 10%, 1%, 0.1%) for an MA(2) model
  • 46. ABC Methods for Bayesian Model Choice Generic ABC model choice Comparison of distance impact 4 1.5 3 1.0 2 0.5 1 0.0 0 0.0 0.2 0.4 0.6 0.8 −2.0 −1.0 0.0 0.5 1.0 1.5 θ1 θ2 Evaluation of the tolerance on the ABC sample against both distances ( = 100%, 10%, 1%, 0.1%) for an MA(2) model
  • 47. ABC Methods for Bayesian Model Choice Generic ABC model choice MA(q) divergence 1.0 1.0 1.0 1.0 0.8 0.8 0.8 0.8 0.6 0.6 0.6 0.6 0.4 0.4 0.4 0.4 0.2 0.2 0.2 0.2 0.0 0.0 0.0 0.0 1 2 1 2 1 2 1 2 Evolution [against ] of ABC Bayes factor, in terms of frequencies of visits to models MA(1) (left) and MA(2) (right) when equal to 10, 1, .1, .01% quantiles on insufficient autocovariance distances. Sample of 50 points from a MA(2) with θ1 = 0.6, θ2 = 0.2. True Bayes factor equal to 17.71.
  • 48. ABC Methods for Bayesian Model Choice Generic ABC model choice MA(q) divergence 1.0 1.0 1.0 1.0 0.8 0.8 0.8 0.8 0.6 0.6 0.6 0.6 0.4 0.4 0.4 0.4 0.2 0.2 0.2 0.2 0.0 0.0 0.0 0.0 1 2 1 2 1 2 1 2 Evolution [against ] of ABC Bayes factor, in terms of frequencies of visits to models MA(1) (left) and MA(2) (right) when equal to 10, 1, .1, .01% quantiles on insufficient autocovariance distances. Sample of 50 points from a MA(1) model with θ1 = 0.6. True Bayes factor B21 equal to .004.
  • 49. ABC Methods for Bayesian Model Choice Generic ABC model choice Further comments ‘There should be the possibility that for the same model, but different (non-minimal) [summary] statistics (so ∗ different η’s: η1 and η1 ) the ratio of evidences may no longer be equal to one.’ [Michael Stumpf, Jan. 28, 2011, ’Og] Using different summary statistics [on different models] may indicate the loss of information brought by each set but agreement does not lead to trustworthy approximations.
  • 50. ABC Methods for Bayesian Model Choice Generic ABC model choice A population genetics evaluation Population genetics example with 3 populations 2 scenari 15 individuals 5 loci single mutation parameter
  • 51. ABC Methods for Bayesian Model Choice Generic ABC model choice A population genetics evaluation Population genetics example with 3 populations 2 scenari 15 individuals 5 loci single mutation parameter 24 summary statistics 2 million ABC proposal importance [tree] sampling alternative
  • 52. ABC Methods for Bayesian Model Choice Generic ABC model choice A population genetics evaluation Population genetics example with 3 populations 2 scenari 15 individuals 5 loci single mutation parameter 24 summary statistics 2 million ABC proposal importance [tree] sampling alternative
  • 53. ABC Methods for Bayesian Model Choice Generic ABC model choice Stability of importance sampling 1.0 1.0 1.0 1.0 1.0 q q 0.8 0.8 0.8 0.8 0.8 0.6 0.6 0.6 0.6 0.6 q 0.4 0.4 0.4 0.4 0.4 0.2 0.2 0.2 0.2 0.2 0.0 0.0 0.0 0.0 0.0
  • 54. ABC Methods for Bayesian Model Choice Generic ABC model choice Comparison with ABC Use of 24 summary statistics and DIY-ABC logistic correction 1.0 qq qq qq qq qqq qq q q qq q q qq q q q q q q q q q q qq q q qq q q q qq qq qq q q q q qq q q q q q q q q q q q qq q q q qq q q 0.8 q q q q q q q q q q q q qq q q q q q q q qq q q q q q qqq q q qq q ABC direct and logistic q q qq q q q q q q q 0.6 q q q qq q q qq q q q q q q q q q q q q q q q q q q q q q q q q q qq qq q q qq 0.4 qq q q q q q q q q q q q q q q q q q 0.2 q qq q q q q q q q q q q 0.0 qq 0.0 0.2 0.4 0.6 0.8 1.0 importance sampling
  • 55. ABC Methods for Bayesian Model Choice Generic ABC model choice Comparison with ABC Use of 15 summary statistics and DIY-ABC logistic correction 6 4 q q q q q q q qq 2 q q q q qq ABC direct q q q qq q qq q q q q qq qq q qq q q q q q qq q q q q q qq q q qq q qq q q qq q q q qqq q q q q q q q q 0 qq qq q q q q qq qq q q qq q q q q q q q q −2 −4 −4 −2 0 2 4 6 importance sampling
  • 56. ABC Methods for Bayesian Model Choice Generic ABC model choice Comparison with ABC Use of 15 summary statistics and DIY-ABC logistic correction qq 6 q q q q q q q q q q q 4 q qq q qq qq q q q q q q q q q q qq q q q q q q q ABC direct and logistic q q q q q q q qq q q q q q q q qq q 2 q q q q q qq q qq q q q qq qq q qq qq q q q qq q q qq qq qq q q qqqq qq q q q qqqqq qq q q q qq q q q q qq q qq q q qqq q q qqqqq q q qq q q qq q q q qq 0 qq qqq q q qq q qqq q q q q q q q q q q q q q q q q q q q q qq q q −2 q q q q q −4 q q q −4 −2 0 2 4 6 importance sampling
  • 57. ABC Methods for Bayesian Model Choice Generic ABC model choice Comparison with ABC Use of 15 summary statistics and DIY-ABC logistic correction q 2 q q 1 q q q q q q q q q q log−ratio q q q q q qq q q q qqq q q q q q q q qq q q q q q qq qq q qq q q qqqqqq q q qqqq qq qq q qq qq q qq q qq qq q q q qq q qq q q qqq q 0 q q q q qq q q q qq q q q q q q qq q qq q qq qq q qq qqq qq qq q q q q qq q qqqq q q q q q q q q qq qqq q qq qq q q qq q q q q qq q q qq q q q q q q q q qq qqqq q q q q q q q q q −1 q q q q −2 q 0 20 40 60 80 100 index
  • 58. ABC Methods for Bayesian Model Choice Generic ABC model choice A second population genetics experiment three populations, two divergent 100 gen. ago two scenarios [third pop. recent admixture between first two pop. / diverging from pop. 1 5 gen. ago] In scenario 1, admixture rate 0.7 from pop. 1 100 datasets with 100 diploid individuals per population, 50 independent microsatellite loci. Effective population size of 1000 and mutation rates of 0.0005. 6 parameters: admixture rate (U [0.1, 0.9]), three effective population sizes (U [200, 2000]), the time of admixture/second divergence (U [1, 10]) and time of first divergence (U [50, 500]).
  • 59. ABC Methods for Bayesian Model Choice Generic ABC model choice Results IS algorithm performed with 100 coalescent trees per particle and 50,000 particles [12 calendar days using 376 processors] Using ten times as many loci and seven times as many individuals degrades the confidence in the importance sampling approximation because of an increased variability in the likelihood. 0.9 0.8 q 0.7 0.6 q 0.5
  • 60. ABC Methods for Bayesian Model Choice Generic ABC model choice Results Blurs potential divergence between ABC and genuine posterior probabilities because both are overwhelmingly close to one, due to the high information content of the data. 1.0 q qq q qq q q qqq qq q q qq qq q q q qq qq q qq q q q q q 0.8 q Importance Sampling estimates of q P(M=1|y) with 1,000 particles q q 0.6 0.4 0.2 0.0 0.0 0.2 0.4 0.6 0.8 1.0 Importance Sampling estimates of
  • 61. ABC Methods for Bayesian Model Choice Generic ABC model choice Results Blurs potential divergence between ABC and genuine posterior probabilities because both are overwhelmingly close to one, due to the high information content of the data. 1.0 q q qqq q qqqq q q qqqqq q qqq qq qq qq qq qq q q q qq qqq Importance Sampling estimates of P(M=1|y) with 50,000 particles q q qq q q q qq q q q q q 0.8 q q q q 0.6 0.4 0.2 0.0 0.0 0.2 0.4 0.6 0.8 1.0
  • 62. ABC Methods for Bayesian Model Choice Generic ABC model choice The only safe cases Besides specific models like Gibbs random fields, using distances over the data itself escapes the discrepancy... [Toni & Stumpf, 2010;Sousa et al., 2009]
  • 63. ABC Methods for Bayesian Model Choice Generic ABC model choice The only safe cases Besides specific models like Gibbs random fields, using distances over the data itself escapes the discrepancy... [Toni & Stumpf, 2010;Sousa et al., 2009] ...and so does the use of more informal model fitting measures [Ratmann, Andrieu, Richardson and Wiujf, 2009]