SlideShare a Scribd company logo
1 of 54
Download to read offline
Unbiased Markov chain Monte Carlo methods
Pierre E. Jacob
Department of Statistics, Harvard University
joint work with John O’Leary, Yves Atchad´e, and other fantastic
people acknowledged throughout
Department of Statistics, Stanford University
April 16, 2019
Pierre E. Jacob Unbiased MCMC
Outline
1 Monte Carlo and bias
2 Unbiased estimators from Markov kernels
3 Numerical experiments
4 Beyond parallel computation
Pierre E. Jacob Unbiased MCMC
Outline
1 Monte Carlo and bias
2 Unbiased estimators from Markov kernels
3 Numerical experiments
4 Beyond parallel computation
Pierre E. Jacob Unbiased MCMC
Setting
Continuous or discrete space of dimension d.
Target probability distribution π,
with probability density function x → π(x).
Goal: approximate π, e.g. approximate
Eπ[h(X)] = h(x)π(x)dx = π(h),
for a class of “test” functions h.
Pierre E. Jacob Unbiased MCMC
Markov chain Monte Carlo
Initially, X0 ∼ π0, then Xt|Xt−1 ∼ P(Xt−1, ·) for t = 1, . . . , T.
Estimator:
1
T − b
T
t=b+1
h(Xt),
where b iterations are discarded as burn-in.
Might converge to Eπ[h(X)] as T → ∞ by the ergodic theorem.
Biased for any fixed b, T, since π0 = π.
Averaging independent copies of such estimators for fixed T, b
would not provide a consistent estimator of Eπ[h(X)]
as the number of independent copies goes to infinity.
Pierre E. Jacob Unbiased MCMC
Example: Metropolis–Hastings kernel P
Initialize: X0 ∼ π0.
At each iteration t ≥ 0, with Markov chain at state Xt,
1 propose X ∼ k(Xt, ·),
2 sample U ∼ U([0, 1]),
3 if
U ≤
π(X )k(X , Xt)
π(Xt)k(Xt, X )
,
set Xt+1 = X , otherwise set Xt+1 = Xt.
Hastings, Monte Carlo sampling methods using Markov chains and
their applications, Biometrika, 1970.
Pierre E. Jacob Unbiased MCMC
MCMC path
π = N(0, 1), MH with Normal proposal std = 0.5, π0 = N(10, 32
)
Pierre E. Jacob Unbiased MCMC
MCMC marginal distributions
π = N(0, 1), RWMH with Normal proposal std = 0.5, π0 = N(10, 32
)
Pierre E. Jacob Unbiased MCMC
Sequential Monte Carlo samplers
π = N(0, 1), adaptive SMC sampler with MH moves, π0 = N(10, 32
)
Neal, Annealed importance sampling, STCO, 2001.
Del Moral, Doucet, Jasra, SMC samplers, JRSS B, 2006.
Pierre E. Jacob Unbiased MCMC
Empirical measure approximations
MCMC: ˆπ = 1
T
T
t=1 δXt → π as T → ∞.
SMC: ˆπ = N
n=1 ωnδXn → π as N → ∞.
Proposed: ˆπ = N
n=1 ωnδXn , with E[Eˆπ[h(X)]] = Eπ[h(X)].
Pierre E. Jacob Unbiased MCMC
Do we care about unbiased estimators?
In classical point estimation, unbiasedness is not crucial.
Larry Wasserman in “All of Statistics” (2003) writes:
Unbiasedness used to receive much attention but these
days is considered less important.
On the other hand, Jeff Rosenthal in “Parallel computing and
Monte Carlo algorithms” (2000) writes:
When running parallel Monte Carlo with many
computers, it is more important to start with an
unbiased (or low-bias) estimate than with a
low-variance estimate.
Pierre E. Jacob Unbiased MCMC
Parallel computing
R processors generate unbiased estimators, independently.
The number of estimators completed by time t is random.
There are different ways of aggregating estimators,
either discarding on-going calculations at time t or not.
Different regimes can be considered, e.g.
as R → ∞ for fixed t,
as t → ∞ for fixed R,
as t and R both go to infinity.
Glynn & Heidelberger, Analysis of parallel replicated
simulations under a completion time constraint, 1991.
Pierre E. Jacob Unbiased MCMC
Parallel computing
Pierre E. Jacob Unbiased MCMC
Outline
1 Monte Carlo and bias
2 Unbiased estimators from Markov kernels
3 Numerical experiments
4 Beyond parallel computation
Pierre E. Jacob Unbiased MCMC
Coupled chains
Generate two Markov chains (Xt) and (Yt) as follows,
sample X0 and Y0 from π0 (independently or not),
sample X1|X0 ∼ P(X0, ·),
for t ≥ 1, sample (Xt+1, Yt)|(Xt, Yt−1) ∼ ¯P ((Xt, Yt−1), ·),
¯P is such that Xt+1|Xt ∼ P(Xt, ·) and Yt|Yt−1 ∼ P(Yt−1, ·),
so each chain marginally evolves according to P.
⇒ Xt and Yt have the same distribution for all t ≥ 0.
¯P is also such that ∃τ such that Xτ = Yτ−1,
and the chains are faithful.
Pierre E. Jacob Unbiased MCMC
Debiasing idea (one slide version)
Limit as a telescopic sum, for all k ≥ 0,
Eπ[h(X)] = lim
t→∞
E[h(Xt)] = E[h(Xk)] +
∞
t=k+1
E[h(Xt) − h(Xt−1)].
Since for all t ≥ 0, Xt and Yt have the same distribution,
= E[h(Xk)] +
∞
t=k+1
E[h(Xt) − h(Yt−1)].
If we can swap expectation and limit,
= E[h(Xk) +
∞
t=k+1
(h(Xt) − h(Yt−1))],
so h(Xk) + ∞
t=k+1(h(Xt) − h(Yt−1)) is unbiased.
Pierre E. Jacob Unbiased MCMC
Unbiased estimators
Unbiased estimator is given by
Hk(X, Y ) = h(Xk) +
τ−1
t=k+1
(h(Xt) − h(Yt−1)),
with the convention τ−1
t=k+1{·} = 0 if τ − 1 < k + 1.
h(Xk) alone is biased; the other terms correct for the bias.
Cost: τ − 1 calls to ¯P and 1 + max(0, k − τ) calls to P.
Glynn & Rhee, Exact estimation for Markov chain equilibrium
expectations, 2014.
Pierre E. Jacob Unbiased MCMC
Conditions
Jacob, O’Leary, Atchad´e, Unbiased MCMC with couplings.
1 Marginal chain converges:
E[h(Xt)] → Eπ[h(X)],
and h(Xt) has (2 + η)-finite moments for all t.
2 Meeting time τ has geometric tails:
∃C < +∞ ∃δ ∈ (0, 1) ∀t ≥ 0 P(τ > t) ≤ Cδt
.
3 Chains stay together: Xt = Yt−1 for all t ≥ τ.
Condition 2 itself implied by e.g. geometric drift condition.
Under these conditions, Hk(X, Y ) is unbiased, has finite
expected cost and finite variance, for all k.
Pierre E. Jacob Unbiased MCMC
Conditions: update
Middleton, Deligiannidis, Doucet, Jacob, Unbiased MCMC for
intractable target distributions.
1 Marginal chain converges:
E[h(Xt)] → Eπ[h(X)],
and h(Xt) has (2 + η)-finite moments for all t.
2 Meeting time τ has polynomial tails:
∃C < +∞ ∃δ > 2(2η−1
+ 1) ∀t ≥ 0 P(τ > t) ≤ Ct−δ
.
3 Chains stay together: Xt = Yt−1 for all t ≥ τ.
Condition 2 itself implied by e.g. polynomial drift condition.
Pierre E. Jacob Unbiased MCMC
Improved unbiased estimators
Efficiency matters, thus in practice we recommend a variation
of the previous estimator, defined for integers k ≤ m as
Hk:m(X, Y ) =
1
m − k + 1
m
t=k
Ht(X, Y )
which can also be written
1
m − k + 1
m
t=k
h(Xt)+
τ−1
t=k+1
min 1,
t − k
m − k + 1
(h(Xt)−h(Yt−1)),
i.e. standard MCMC average + bias correction term.
As k → ∞, bias correction is zero with increasing probability.
Pierre E. Jacob Unbiased MCMC
Efficiency
Writing Hk:m(X, Y ) = MCMCk:m + BCk:m,
then, denoting the MSE of MCMCk:m by MSEk:m,
V[Hk:m(X, Y )] ≤ MSEk:m+2 MSEk:m E BC2
k:m +E BC2
k:m .
Under geometric drift condition and regularity assumptions on
h, for some δ < 1, C < ∞,
E[BC2
k:m] ≤
Cδk
(m − k + 1)2
,
and δ is directly related to tails of the meeting time.
Similarly under polynomial drift, we obtain a term of the form
k−δ in the numerator instead of δk.
Pierre E. Jacob Unbiased MCMC
Signed measure estimator
Replacing function evaluations by delta masses leads to
ˆπ(·) =
1
m − k + 1
m
t=k
δXt (·)
+
τ−1
t=k+1
min 1,
t − k
m − k + 1
(δXt (·) − δYt−1 (·)),
which is of the form
ˆπ(·) =
N
n=1
ωnδXn (·),
where N
n=1 ωn = 1 but some ωn might be negative.
Unbiasedness reads: E[Eˆπ[h(X)]] = Eπ[h(X)].
Pierre E. Jacob Unbiased MCMC
Determining a warm-up period
Total variation distance between Xk ∼ πk and π = limk→∞ πk:
πk − π TV =
1
2
sup
h:|h|≤1
|E[h(Xk)] − Eπ[h(X)]|
=
1
2
sup
h:|h|≤1
|E[
τ−1
t=k+1
h(Xt) − h(Yt−1)]|
≤ E[max(0, τ − k + 1)].
0.00
0.01
0.02
0.03
0.04
0 50 100 150 200
meeting time
density
1e−03
1e−02
1e−01
1e+00
1e+01
0 50 100
k
upperbound
Pierre E. Jacob Unbiased MCMC
Designing coupled chains
To implement the proposed unbiased estimators,
we need to sample from a Markov kernel ¯P,
such that, when (Xt+1, Yt) is sampled from ¯P ((Xt, Yt−1), ·),
marginally Xt+1|Xt ∼ P(Xt, ·), and Yt|Yt−1 ∼ P(Yt−1, ·),
it is possible that Xt+1 = Yt exactly for some t ≥ 0,
if Xt = Yt−1, then Xt+1 = Yt almost surely.
Pierre E. Jacob Unbiased MCMC
Couplings of MCMC algorithms
Many practical couplings in the literature. . .
Propp & Wilson, Exact sampling with coupled Markov chains
and applications to statistical mechanics, Random Structures &
Algorithms, 1996.
Johnson, Studying convergence of Markov chain Monte Carlo
algorithms using coupled sample paths, JASA, 1996.
Neal, Circularly-coupled Markov chain sampling, UoT tech
report, 1999.
Pinto & Neal, Improving Markov chain Monte Carlo estimators
by coupling to an approximating chain, UoT tech report, 2001.
Glynn & Rhee, Exact estimation for Markov chain equilibrium
expectations, Journal of Applied Probability, 2014.
Pierre E. Jacob Unbiased MCMC
Couplings of MCMC algorithms: contributions
Conditional particle filters
Jacob, Lindsten, Sch¨on, Smoothing with Couplings of
Conditional Particle Filters, JASA, 2018.
Metropolis–Hastings, Gibbs samplers, parallel tempering
Jacob, O’Leary, Atchad´e, Unbiased MCMC with couplings,
submitted.
Hamiltonian Monte Carlo
Heng & Jacob, Unbiased HMC with couplings, Biometrika, 2019.
Pseudo-marginal MCMC, exchange algorithm
Middleton, Deligiannidis, Doucet, Jacob, Unbiased MCMC for
intractable target distributions, submitted.
Particle independent Metropolis–Hastings
Middleton, Deligiannidis, Doucet, Jacob, Unbiased Smoothing
using Particle Independent Metropolis-Hastings, AISTATS, 2019.
Pierre E. Jacob Unbiased MCMC
Couplings
(X, Y ) follows a coupling of p and q if X ∼ p and Y ∼ q.
The coupling inequality states that
P(X = Y ) ≤ 1 − p − q TV,
for any coupling, with p − q TV = 1
2 |p(x) − q(x)|dx.
Maximal couplings achieve the bound.
Pierre E. Jacob Unbiased MCMC
Maximal coupling of Gamma and Normal
Pierre E. Jacob Unbiased MCMC
Maximal coupling: sampling algorithm
Requires: evaluations of p and q, sampling from p and q.
1 Sample X ∼ p and W ∼ U([0, 1]).
If W ≤ q(X)/p(X), output (X, X).
2 Otherwise, sample Y ∼ q and W ∼ U([0, 1])
until W > p(Y )/q(Y ), and output (X, Y ).
Output: a pair (X, Y ) such that X ∼ p, Y ∼ q
and P(X = Y ) is maximal.
Pierre E. Jacob Unbiased MCMC
Back to Metropolis–Hastings (kernel P)
At each iteration t, Markov chain at state Xt,
1 propose X ∼ k(Xt, ·),
2 sample U ∼ U([0, 1]),
3 if
U ≤
π(X )k(X , Xt)
π(Xt)k(Xt, X )
,
set Xt+1 = X , otherwise set Xt+1 = Xt.
How to propagate two MH chains from states Xt and Yt−1
such that {Xt+1 = Yt} can happen?
Pierre E. Jacob Unbiased MCMC
Coupling of Metropolis–Hastings (kernel ¯P)
At each iteration t, two Markov chains at states Xt, Yt−1,
1 propose (X , Y ) from max coupling of k(Xt, ·), k(Yt−1, ·),
2 sample U ∼ U([0, 1]),
3 if
U ≤
π(X )k(X , Xt)
π(Xt)k(Xt, X )
,
set Xt+1 = X , otherwise set Xt+1 = Xt,
if
U ≤
π(Y )k(Y , Yt−1)
π(Yt−1)k(Yt−1, Y )
,
set Yt = Y , otherwise set Yt = Yt−1.
Pierre E. Jacob Unbiased MCMC
Coupling of Gibbs
Gibbs sampler: update component i of the chain,
leaving π(dxi|x1, . . . , xi−1, xi+1, . . . , xd) invariant.
For instance, we can propose X ∼ k(Xi
t, ·) to replace Xi
t,
and accept or not with a Metropolis–Hastings step.
These proposals can be maximally coupled across two chains,
at each component update.
The chains meet when all components have met.
Likewise we can couple parallel tempering chains,
and meeting occurs when entire ensembles of chains meet.
Pierre E. Jacob Unbiased MCMC
Hamiltonian Monte Carlo
Introduce potential energy U(q) = − log π(q),
and total energy E(q, p) = U(q) + 1
2|p|2.
Hamiltonian dynamics for (q(s), p(s)), where s ≥ 0:
d
ds
q(s) = pE(q(s), p(s)),
d
ds
p(s) = − qE(q(s), p(s)).
Solving Hamiltonian dynamics exactly is not feasible,
but discretization + Metropolis–Hastings correction ensure that
π remains invariant.
Common random numbers can make two HMC chains contract,
under assumptions on the target such as strong log-concavity.
Heng & Jacob, Unbiased HMC with couplings, Biometrika, 2019.
Pierre E. Jacob Unbiased MCMC
Coupling of HMC
Figure 2 of Mangoubi & Smith, Rapid mixing of HMC strongly
log-concave distributions, 2017.
Coupling two copies X1, X2, . . . (blue) and Y1, Y2, . . .
(green) of HMC by choosing same momentum pi at
every step.
Pierre E. Jacob Unbiased MCMC
Particle Independent Metropolis–Hastings
Initialization: run SMC, obtain ˆπ(0)(·), ˆZ(0),
where E[ ˆZ(0)] = Z, the normalizing constant of π.
At each iteration t, given (ˆπ(t)(·), ˆZ(t)),
1 run SMC, obtain (π (·), Z ),
2 sample U ∼ U([0, 1]),
3 if
U ≤
Z
ˆZ(t)
,
set (t + 1)-th state to (π (·), Z ),
otherwise set (t + 1)-th state to (ˆπ(t)(·), ˆZ(t)).
For any number of particles, T−1 T
t=1 ˆπ(t)(·) →T→∞ π.
Andrieu, Doucet, Holenstein, Particle MCMC, JRSS B, 2010.
Pierre E. Jacob Unbiased MCMC
Coupled Particle Independent Metropolis–Hastings
At each iteration t, given (ˆπ(t)(·), ˆZ(t)), and (˜π(t−1)(·), ˜Z(t−1)),
1 run SMC, obtain (π (·), Z ),
2 sample U ∼ U([0, 1]),
3 if
U ≤
Z
ˆZ(t)
,
set (t + 1)-th “1st” state to (π (·), Z ),
otherwise set (t + 1)-th “1st” state to (ˆπ(t)(·), ˆZ(t)).
4 if
U ≤
Z
˜Z(t−1)
,
set t-th “2nd” state to (π (·), Z ),
otherwise set t-th “2nd” state to (˜π(t−1)(·), ˜Z(t−1)).
Pierre E. Jacob Unbiased MCMC
Unbiased SMC
PIMH turns any SMC sampler into an MCMC scheme
that can be debiased using the coupled chains machinery.
Thus any MCMC algorithm that can be plugged in an
SMC sampler, can then be subsequently de-biased.
No need for couplings of the MCMC chains themselves.
Middleton, Deligiannidis, Doucet, Jacob, Unbiased Smoothing using
Particle Independent Metropolis-Hastings, AISTATS, 2019.
Pierre E. Jacob Unbiased MCMC
Outline
1 Monte Carlo and bias
2 Unbiased estimators from Markov kernels
3 Numerical experiments
4 Beyond parallel computation
Pierre E. Jacob Unbiased MCMC
Linear regression: model
Johndrow, Orenstein, Bhattacharya, Bayes Shrinkage at GWAS
scale: Convergence and Approximation Theory of a Scalable
MCMC Algorithm for the Horseshoe Prior.
∀i ∈ {1, . . . , n} Yi|Xi, β, σ2
∼ N(XT
i β, σ2
)
∀j ∈ {1, . . . , p} βj|σ2
, ηj, ξ ∼ N(0,
σ2
ξ
η−1
j ), η
−1/2
j ∼ C+
(0, 1)
ξ−1/2
∼ C+
(0, 1), σ2
∼ InvGamma(
1
2
,
1
2
)
Data generated with Xij
i.i.d.
∼ N(0, 1), σ2 = 0.12,
βj = 4 for j = 1, . . . , 5,
βj = 2(6−j)/2 for j = 6, . . . , 20, βj = 0 for j ≥ 21.
Pierre E. Jacob Unbiased MCMC
Linear regression: Gibbs sampler
η|β, ξ, σ2 from distribution
∝ p
j=1
1
1+ηj
exp(−
β2
j ξ
2σ2 ηj)1(ηj > 0),
using rejection sampler for each component j,
⇒ maximal coupling for each component j,
ξ|η using Metropolis–Hastings step,
⇒ maximal coupling of the Normal proposals,
σ2|η, ξ from Inverse Gamma distribution,
⇒ maximal coupling,
β|η, ξ, σ2 from p-dimensional Normal distribution,
⇒ common random numbers in fast algorithm from
Bhattacharya, Chakraborty, Mallick (2016).
Pierre E. Jacob Unbiased MCMC
Linear regression: meeting times
p=100 p=250 p=500
n=100n=250n=500
10 100 1000 10000 10 100 1000 10000 10 100 1000 10000
0
5
10
15
20
0
5
10
15
20
0
5
10
15
20
meeting time
count
Biswas & Jacob, Unbiased MCMC for linear regressions.
Pierre E. Jacob Unbiased MCMC
Linear regression: meeting times
p=100 p=250 p=500
n=100n=250n=500
1 10 100 100010000 1 10 100 100010000 1 10 100 100010000
0
10
20
0
10
20
0
10
20
duration (seconds)
count
Biswas & Jacob, Unbiased MCMC for linear regressions.
Pierre E. Jacob Unbiased MCMC
Whisker experiment
Activation of neurons of rats, as their whiskers are being moved
with a periodic stimulus.
The activation of a neuron is recorded as a binary variable for
each time and each experiment. These activation variables are
then aggregated by summing over the M experiments at each
time step.
Data kindly shared by Demba Ba (SEAS).
Lin, Zhang, Heng, Allsop, Tye, Jacob, Ba, Clustering Time Series
with Nonlinear Dynamics: A Bayesian Non-Parametric and
Particle-Based Approach, AISTATS 2019.
Pierre E. Jacob Unbiased MCMC
Whisker experiment: data
Model: X0 ∼ N(0, 1), Xt|Xt−1 ∼ N(aXt−1, σ2
X),
Yt|Xt ∼ Binom(M, (1 + exp(−Xt))−1
).
Pierre E. Jacob Unbiased MCMC
Whisker experiment: meeting times
Middleton, Deligiannidis, Doucet, Jacob, Unbiased MCMC for
intractable target distributions.
Heng, Bishop, Deligiannidis & Doucet, Controlled SMC.
Pierre E. Jacob Unbiased MCMC
Whisker experiment: in parallel
Computation chronology, on 100 processors over 23 hours, with
k = 1, 000, m = 10k.
Pierre E. Jacob Unbiased MCMC
Whisker experiment: results
Approximations of marginal posteriors, k = 1, 000, m = 10k.
Red line corresponds to PMMH run (250,000 iterations).
Pierre E. Jacob Unbiased MCMC
Outline
1 Monte Carlo and bias
2 Unbiased estimators from Markov kernels
3 Numerical experiments
4 Beyond parallel computation
Pierre E. Jacob Unbiased MCMC
Unbiased property
Unbiased MCMC estimators can be used to tackle problems for
which standard MCMC methods are ill-suited.
For instance, problems where one needs to approximate
h(x1, x2)π2(x2|x1)dx2 π1(x1)dx1,
or equivalently
E1 [E2[h(X1, X2)|X1]] .
Pierre E. Jacob Unbiased MCMC
Example 1: cut distribution
First model, parameter θ1, data Y1, posterior π1(θ1|Y1).
Second model, parameter θ2, data Y2, conditional posterior:
π2(θ2|Y2, θ1) =
p2(θ2|θ1)p2(Y2|θ1, θ2)
p2(Y2|θ1)
.
Plummer, Cuts in Bayesian graphical models, 2015.
Jacob, Murray, Holmes, Robert, Better together? Statistical learning
in models made of modules.
Pierre E. Jacob Unbiased MCMC
Example 2: normalizing constants
We might be interested in normalizing constant Z of
π(x) ∝ exp(−U(x)), defined as Z = exp(−U(x))dx.
Introduce
∀λ ∈ [0, 1] πλ(x) ∝ exp(−Uλ(x)).
Thermodynamics integration or path sampling identity:
log
Z1
Z0
= −
1
0
λUλ(x)πλ(dx)
q(λ)
q(dλ),
where q(dλ) is any distribution supported on [0, 1].
Rischard, Jacob, Pillai, Unbiased estimation of log normalizing
constants with applications to Bayesian cross-validation.
Pierre E. Jacob Unbiased MCMC
Example 3: Bayesian cross-validation
Data y = {y1, . . . , yn}, made of n units.
Partition y into training T and validation V sets.
For each split y = {T, V }, assess predictive performance with
log p(V |T) = log p(V |θ, T)p(dθ|T).
Note, log p(V |T) is a log-ratio of normalizing constants.
Finally we might want to approximate
CV =
n
nT
−1
T,V
log p(V |T).
Rischard, Jacob, Pillai, Unbiased estimation of log normalizing
constants with applications to Bayesian cross-validation.
Pierre E. Jacob Unbiased MCMC
Discussion
Perfect samplers, designed to sample i.i.d. from π, would
yield the same benefits and more.
What about regeneration?
Implications of lack of bias are numerous.
Proposed estimators can be obtained
using coupled MCMC kernels,
or using the “MCMC → SMC → coupled PIMH” route.
Cannot work if underlying MCMC doesn’t work. . .
but parallelism allows for more expensive MCMC kernels.
Thank you for listening, and special thanks to James Johndrow!
Pierre E. Jacob Unbiased MCMC
References
with Fredrik Lindsten, Thomas Sch¨on
Smoothing with Couplings of Conditional Particle Filters, 2018.
with John O’Leary, Yves F. Atchad´e
Unbiased Markov chain Monte Carlo with couplings, 2019.
with Jeremy Heng
Unbiased Hamiltonian Monte Carlo with couplings, 2019.
with Lawrence Middleton, George Deligiannidis, Arnaud
Doucet
Unbiased Markov chain Monte Carlo for intractable target
distributions, 2019.
Unbiased Smoothing using Particle Independent
Metropolis-Hastings, 2019.
with Maxime Rischard, Natesh Pillai
Unbiased estimation of log normalizing constants with
applications to Bayesian cross-validation, 2019.
Pierre E. Jacob Unbiased MCMC

More Related Content

What's hot

Bayesian inversion of deterministic dynamic causal models
Bayesian inversion of deterministic dynamic causal modelsBayesian inversion of deterministic dynamic causal models
Bayesian inversion of deterministic dynamic causal modelskhbrodersen
 
Rao-Blackwellisation schemes for accelerating Metropolis-Hastings algorithms
Rao-Blackwellisation schemes for accelerating Metropolis-Hastings algorithmsRao-Blackwellisation schemes for accelerating Metropolis-Hastings algorithms
Rao-Blackwellisation schemes for accelerating Metropolis-Hastings algorithmsChristian Robert
 
Unbiased Bayes for Big Data
Unbiased Bayes for Big DataUnbiased Bayes for Big Data
Unbiased Bayes for Big DataChristian Robert
 
Complexity of exact solutions of many body systems: nonequilibrium steady sta...
Complexity of exact solutions of many body systems: nonequilibrium steady sta...Complexity of exact solutions of many body systems: nonequilibrium steady sta...
Complexity of exact solutions of many body systems: nonequilibrium steady sta...Lake Como School of Advanced Studies
 
Neutral Electronic Excitations: a Many-body approach to the optical absorptio...
Neutral Electronic Excitations: a Many-body approach to the optical absorptio...Neutral Electronic Excitations: a Many-body approach to the optical absorptio...
Neutral Electronic Excitations: a Many-body approach to the optical absorptio...Claudio Attaccalite
 
Introduction to MCMC methods
Introduction to MCMC methodsIntroduction to MCMC methods
Introduction to MCMC methodsChristian Robert
 
Omiros' talk on the Bernoulli factory problem
Omiros' talk on the  Bernoulli factory problemOmiros' talk on the  Bernoulli factory problem
Omiros' talk on the Bernoulli factory problemBigMC
 
MCMC and likelihood-free methods
MCMC and likelihood-free methodsMCMC and likelihood-free methods
MCMC and likelihood-free methodsChristian Robert
 
Introduction to Diffusion Monte Carlo
Introduction to Diffusion Monte CarloIntroduction to Diffusion Monte Carlo
Introduction to Diffusion Monte CarloClaudio Attaccalite
 
Speeding up the Gillespie algorithm
Speeding up the Gillespie algorithmSpeeding up the Gillespie algorithm
Speeding up the Gillespie algorithmColin Gillespie
 
Bayesian regression models and treed Gaussian process models
Bayesian regression models and treed Gaussian process modelsBayesian regression models and treed Gaussian process models
Bayesian regression models and treed Gaussian process modelsTommaso Rigon
 

What's hot (20)

Bayesian inversion of deterministic dynamic causal models
Bayesian inversion of deterministic dynamic causal modelsBayesian inversion of deterministic dynamic causal models
Bayesian inversion of deterministic dynamic causal models
 
Rao-Blackwellisation schemes for accelerating Metropolis-Hastings algorithms
Rao-Blackwellisation schemes for accelerating Metropolis-Hastings algorithmsRao-Blackwellisation schemes for accelerating Metropolis-Hastings algorithms
Rao-Blackwellisation schemes for accelerating Metropolis-Hastings algorithms
 
Chris Sherlock's slides
Chris Sherlock's slidesChris Sherlock's slides
Chris Sherlock's slides
 
Lesage
LesageLesage
Lesage
 
QMC Program: Trends and Advances in Monte Carlo Sampling Algorithms Workshop,...
QMC Program: Trends and Advances in Monte Carlo Sampling Algorithms Workshop,...QMC Program: Trends and Advances in Monte Carlo Sampling Algorithms Workshop,...
QMC Program: Trends and Advances in Monte Carlo Sampling Algorithms Workshop,...
 
QMC Program: Trends and Advances in Monte Carlo Sampling Algorithms Workshop,...
QMC Program: Trends and Advances in Monte Carlo Sampling Algorithms Workshop,...QMC Program: Trends and Advances in Monte Carlo Sampling Algorithms Workshop,...
QMC Program: Trends and Advances in Monte Carlo Sampling Algorithms Workshop,...
 
QMC Program: Trends and Advances in Monte Carlo Sampling Algorithms Workshop,...
QMC Program: Trends and Advances in Monte Carlo Sampling Algorithms Workshop,...QMC Program: Trends and Advances in Monte Carlo Sampling Algorithms Workshop,...
QMC Program: Trends and Advances in Monte Carlo Sampling Algorithms Workshop,...
 
Unbiased Bayes for Big Data
Unbiased Bayes for Big DataUnbiased Bayes for Big Data
Unbiased Bayes for Big Data
 
Complexity of exact solutions of many body systems: nonequilibrium steady sta...
Complexity of exact solutions of many body systems: nonequilibrium steady sta...Complexity of exact solutions of many body systems: nonequilibrium steady sta...
Complexity of exact solutions of many body systems: nonequilibrium steady sta...
 
QMC Program: Trends and Advances in Monte Carlo Sampling Algorithms Workshop,...
QMC Program: Trends and Advances in Monte Carlo Sampling Algorithms Workshop,...QMC Program: Trends and Advances in Monte Carlo Sampling Algorithms Workshop,...
QMC Program: Trends and Advances in Monte Carlo Sampling Algorithms Workshop,...
 
Neutral Electronic Excitations: a Many-body approach to the optical absorptio...
Neutral Electronic Excitations: a Many-body approach to the optical absorptio...Neutral Electronic Excitations: a Many-body approach to the optical absorptio...
Neutral Electronic Excitations: a Many-body approach to the optical absorptio...
 
QMC Program: Trends and Advances in Monte Carlo Sampling Algorithms Workshop,...
QMC Program: Trends and Advances in Monte Carlo Sampling Algorithms Workshop,...QMC Program: Trends and Advances in Monte Carlo Sampling Algorithms Workshop,...
QMC Program: Trends and Advances in Monte Carlo Sampling Algorithms Workshop,...
 
Introduction to MCMC methods
Introduction to MCMC methodsIntroduction to MCMC methods
Introduction to MCMC methods
 
Omiros' talk on the Bernoulli factory problem
Omiros' talk on the  Bernoulli factory problemOmiros' talk on the  Bernoulli factory problem
Omiros' talk on the Bernoulli factory problem
 
QMC Program: Trends and Advances in Monte Carlo Sampling Algorithms Workshop,...
QMC Program: Trends and Advances in Monte Carlo Sampling Algorithms Workshop,...QMC Program: Trends and Advances in Monte Carlo Sampling Algorithms Workshop,...
QMC Program: Trends and Advances in Monte Carlo Sampling Algorithms Workshop,...
 
MCMC and likelihood-free methods
MCMC and likelihood-free methodsMCMC and likelihood-free methods
MCMC and likelihood-free methods
 
Introduction to Diffusion Monte Carlo
Introduction to Diffusion Monte CarloIntroduction to Diffusion Monte Carlo
Introduction to Diffusion Monte Carlo
 
Speeding up the Gillespie algorithm
Speeding up the Gillespie algorithmSpeeding up the Gillespie algorithm
Speeding up the Gillespie algorithm
 
Program on Quasi-Monte Carlo and High-Dimensional Sampling Methods for Applie...
Program on Quasi-Monte Carlo and High-Dimensional Sampling Methods for Applie...Program on Quasi-Monte Carlo and High-Dimensional Sampling Methods for Applie...
Program on Quasi-Monte Carlo and High-Dimensional Sampling Methods for Applie...
 
Bayesian regression models and treed Gaussian process models
Bayesian regression models and treed Gaussian process modelsBayesian regression models and treed Gaussian process models
Bayesian regression models and treed Gaussian process models
 

Similar to Unbiased Markov chain Monte Carlo methods

Unbiased Markov chain Monte Carlo
Unbiased Markov chain Monte CarloUnbiased Markov chain Monte Carlo
Unbiased Markov chain Monte CarloJeremyHeng10
 
Unbiased Markov chain Monte Carlo
Unbiased Markov chain Monte CarloUnbiased Markov chain Monte Carlo
Unbiased Markov chain Monte CarloJeremyHeng10
 
Talk at CIRM on Poisson equation and debiasing techniques
Talk at CIRM on Poisson equation and debiasing techniquesTalk at CIRM on Poisson equation and debiasing techniques
Talk at CIRM on Poisson equation and debiasing techniquesPierre Jacob
 
Monte Carlo methods for some not-quite-but-almost Bayesian problems
Monte Carlo methods for some not-quite-but-almost Bayesian problemsMonte Carlo methods for some not-quite-but-almost Bayesian problems
Monte Carlo methods for some not-quite-but-almost Bayesian problemsPierre Jacob
 
Estimation of the score vector and observed information matrix in intractable...
Estimation of the score vector and observed information matrix in intractable...Estimation of the score vector and observed information matrix in intractable...
Estimation of the score vector and observed information matrix in intractable...Pierre Jacob
 
Unbiased Hamiltonian Monte Carlo
Unbiased Hamiltonian Monte CarloUnbiased Hamiltonian Monte Carlo
Unbiased Hamiltonian Monte CarloJeremyHeng10
 
Testing for mixtures by seeking components
Testing for mixtures by seeking componentsTesting for mixtures by seeking components
Testing for mixtures by seeking componentsChristian Robert
 
ABC with data cloning for MLE in state space models
ABC with data cloning for MLE in state space modelsABC with data cloning for MLE in state space models
ABC with data cloning for MLE in state space modelsUmberto Picchini
 
Unbiased Hamiltonian Monte Carlo
Unbiased Hamiltonian Monte Carlo Unbiased Hamiltonian Monte Carlo
Unbiased Hamiltonian Monte Carlo JeremyHeng10
 
Phase-Type Distributions for Finite Interacting Particle Systems
Phase-Type Distributions for Finite Interacting Particle SystemsPhase-Type Distributions for Finite Interacting Particle Systems
Phase-Type Distributions for Finite Interacting Particle SystemsStefan Eng
 
SPDE presentation 2012
SPDE presentation 2012SPDE presentation 2012
SPDE presentation 2012Zheng Mengdi
 
Introduction to Quantum Monte Carlo
Introduction to Quantum Monte CarloIntroduction to Quantum Monte Carlo
Introduction to Quantum Monte CarloClaudio Attaccalite
 
Integration with kernel methods, Transported meshfree methods
Integration with kernel methods, Transported meshfree methodsIntegration with kernel methods, Transported meshfree methods
Integration with kernel methods, Transported meshfree methodsMercier Jean-Marc
 
Computational Information Geometry on Matrix Manifolds (ICTP 2013)
Computational Information Geometry on Matrix Manifolds (ICTP 2013)Computational Information Geometry on Matrix Manifolds (ICTP 2013)
Computational Information Geometry on Matrix Manifolds (ICTP 2013)Frank Nielsen
 
A new implementation of k-MLE for mixture modelling of Wishart distributions
A new implementation of k-MLE for mixture modelling of Wishart distributionsA new implementation of k-MLE for mixture modelling of Wishart distributions
A new implementation of k-MLE for mixture modelling of Wishart distributionsFrank Nielsen
 
Estimation of the score vector and observed information matrix in intractable...
Estimation of the score vector and observed information matrix in intractable...Estimation of the score vector and observed information matrix in intractable...
Estimation of the score vector and observed information matrix in intractable...Pierre Jacob
 
Common Fixed Theorems Using Random Implicit Iterative Schemes
Common Fixed Theorems Using Random Implicit Iterative SchemesCommon Fixed Theorems Using Random Implicit Iterative Schemes
Common Fixed Theorems Using Random Implicit Iterative Schemesinventy
 

Similar to Unbiased Markov chain Monte Carlo methods (20)

Unbiased Markov chain Monte Carlo
Unbiased Markov chain Monte CarloUnbiased Markov chain Monte Carlo
Unbiased Markov chain Monte Carlo
 
Unbiased Markov chain Monte Carlo
Unbiased Markov chain Monte CarloUnbiased Markov chain Monte Carlo
Unbiased Markov chain Monte Carlo
 
Talk at CIRM on Poisson equation and debiasing techniques
Talk at CIRM on Poisson equation and debiasing techniquesTalk at CIRM on Poisson equation and debiasing techniques
Talk at CIRM on Poisson equation and debiasing techniques
 
Monte Carlo methods for some not-quite-but-almost Bayesian problems
Monte Carlo methods for some not-quite-but-almost Bayesian problemsMonte Carlo methods for some not-quite-but-almost Bayesian problems
Monte Carlo methods for some not-quite-but-almost Bayesian problems
 
Estimation of the score vector and observed information matrix in intractable...
Estimation of the score vector and observed information matrix in intractable...Estimation of the score vector and observed information matrix in intractable...
Estimation of the score vector and observed information matrix in intractable...
 
Unbiased Hamiltonian Monte Carlo
Unbiased Hamiltonian Monte CarloUnbiased Hamiltonian Monte Carlo
Unbiased Hamiltonian Monte Carlo
 
Testing for mixtures by seeking components
Testing for mixtures by seeking componentsTesting for mixtures by seeking components
Testing for mixtures by seeking components
 
ABC with data cloning for MLE in state space models
ABC with data cloning for MLE in state space modelsABC with data cloning for MLE in state space models
ABC with data cloning for MLE in state space models
 
Unbiased Hamiltonian Monte Carlo
Unbiased Hamiltonian Monte Carlo Unbiased Hamiltonian Monte Carlo
Unbiased Hamiltonian Monte Carlo
 
Phase-Type Distributions for Finite Interacting Particle Systems
Phase-Type Distributions for Finite Interacting Particle SystemsPhase-Type Distributions for Finite Interacting Particle Systems
Phase-Type Distributions for Finite Interacting Particle Systems
 
SPDE presentation 2012
SPDE presentation 2012SPDE presentation 2012
SPDE presentation 2012
 
Introduction to Quantum Monte Carlo
Introduction to Quantum Monte CarloIntroduction to Quantum Monte Carlo
Introduction to Quantum Monte Carlo
 
Integration with kernel methods, Transported meshfree methods
Integration with kernel methods, Transported meshfree methodsIntegration with kernel methods, Transported meshfree methods
Integration with kernel methods, Transported meshfree methods
 
The Gaussian Hardy-Littlewood Maximal Function
The Gaussian Hardy-Littlewood Maximal FunctionThe Gaussian Hardy-Littlewood Maximal Function
The Gaussian Hardy-Littlewood Maximal Function
 
Computational Information Geometry on Matrix Manifolds (ICTP 2013)
Computational Information Geometry on Matrix Manifolds (ICTP 2013)Computational Information Geometry on Matrix Manifolds (ICTP 2013)
Computational Information Geometry on Matrix Manifolds (ICTP 2013)
 
A new implementation of k-MLE for mixture modelling of Wishart distributions
A new implementation of k-MLE for mixture modelling of Wishart distributionsA new implementation of k-MLE for mixture modelling of Wishart distributions
A new implementation of k-MLE for mixture modelling of Wishart distributions
 
2018 MUMS Fall Course - Statistical Representation of Model Input (EDITED) - ...
2018 MUMS Fall Course - Statistical Representation of Model Input (EDITED) - ...2018 MUMS Fall Course - Statistical Representation of Model Input (EDITED) - ...
2018 MUMS Fall Course - Statistical Representation of Model Input (EDITED) - ...
 
Estimation of the score vector and observed information matrix in intractable...
Estimation of the score vector and observed information matrix in intractable...Estimation of the score vector and observed information matrix in intractable...
Estimation of the score vector and observed information matrix in intractable...
 
QMC: Transition Workshop - Applying Quasi-Monte Carlo Methods to a Stochastic...
QMC: Transition Workshop - Applying Quasi-Monte Carlo Methods to a Stochastic...QMC: Transition Workshop - Applying Quasi-Monte Carlo Methods to a Stochastic...
QMC: Transition Workshop - Applying Quasi-Monte Carlo Methods to a Stochastic...
 
Common Fixed Theorems Using Random Implicit Iterative Schemes
Common Fixed Theorems Using Random Implicit Iterative SchemesCommon Fixed Theorems Using Random Implicit Iterative Schemes
Common Fixed Theorems Using Random Implicit Iterative Schemes
 

More from Pierre Jacob

ISBA 2022 Susie Bayarri lecture
ISBA 2022 Susie Bayarri lectureISBA 2022 Susie Bayarri lecture
ISBA 2022 Susie Bayarri lecturePierre Jacob
 
Estimation of the score vector and observed information matrix in intractable...
Estimation of the score vector and observed information matrix in intractable...Estimation of the score vector and observed information matrix in intractable...
Estimation of the score vector and observed information matrix in intractable...Pierre Jacob
 
Estimation of the score vector and observed information matrix in intractable...
Estimation of the score vector and observed information matrix in intractable...Estimation of the score vector and observed information matrix in intractable...
Estimation of the score vector and observed information matrix in intractable...Pierre Jacob
 
Current limitations of sequential inference in general hidden Markov models
Current limitations of sequential inference in general hidden Markov modelsCurrent limitations of sequential inference in general hidden Markov models
Current limitations of sequential inference in general hidden Markov modelsPierre Jacob
 
On non-negative unbiased estimators
On non-negative unbiased estimatorsOn non-negative unbiased estimators
On non-negative unbiased estimatorsPierre Jacob
 
Path storage in the particle filter
Path storage in the particle filterPath storage in the particle filter
Path storage in the particle filterPierre Jacob
 
SMC^2: an algorithm for sequential analysis of state-space models
SMC^2: an algorithm for sequential analysis of state-space modelsSMC^2: an algorithm for sequential analysis of state-space models
SMC^2: an algorithm for sequential analysis of state-space modelsPierre Jacob
 
PAWL - GPU meeting @ Warwick
PAWL - GPU meeting @ WarwickPAWL - GPU meeting @ Warwick
PAWL - GPU meeting @ WarwickPierre Jacob
 
Presentation of SMC^2 at BISP7
Presentation of SMC^2 at BISP7Presentation of SMC^2 at BISP7
Presentation of SMC^2 at BISP7Pierre Jacob
 
Presentation MCB seminar 09032011
Presentation MCB seminar 09032011Presentation MCB seminar 09032011
Presentation MCB seminar 09032011Pierre Jacob
 

More from Pierre Jacob (10)

ISBA 2022 Susie Bayarri lecture
ISBA 2022 Susie Bayarri lectureISBA 2022 Susie Bayarri lecture
ISBA 2022 Susie Bayarri lecture
 
Estimation of the score vector and observed information matrix in intractable...
Estimation of the score vector and observed information matrix in intractable...Estimation of the score vector and observed information matrix in intractable...
Estimation of the score vector and observed information matrix in intractable...
 
Estimation of the score vector and observed information matrix in intractable...
Estimation of the score vector and observed information matrix in intractable...Estimation of the score vector and observed information matrix in intractable...
Estimation of the score vector and observed information matrix in intractable...
 
Current limitations of sequential inference in general hidden Markov models
Current limitations of sequential inference in general hidden Markov modelsCurrent limitations of sequential inference in general hidden Markov models
Current limitations of sequential inference in general hidden Markov models
 
On non-negative unbiased estimators
On non-negative unbiased estimatorsOn non-negative unbiased estimators
On non-negative unbiased estimators
 
Path storage in the particle filter
Path storage in the particle filterPath storage in the particle filter
Path storage in the particle filter
 
SMC^2: an algorithm for sequential analysis of state-space models
SMC^2: an algorithm for sequential analysis of state-space modelsSMC^2: an algorithm for sequential analysis of state-space models
SMC^2: an algorithm for sequential analysis of state-space models
 
PAWL - GPU meeting @ Warwick
PAWL - GPU meeting @ WarwickPAWL - GPU meeting @ Warwick
PAWL - GPU meeting @ Warwick
 
Presentation of SMC^2 at BISP7
Presentation of SMC^2 at BISP7Presentation of SMC^2 at BISP7
Presentation of SMC^2 at BISP7
 
Presentation MCB seminar 09032011
Presentation MCB seminar 09032011Presentation MCB seminar 09032011
Presentation MCB seminar 09032011
 

Recently uploaded

G9 Science Q4- Week 1-2 Projectile Motion.ppt
G9 Science Q4- Week 1-2 Projectile Motion.pptG9 Science Q4- Week 1-2 Projectile Motion.ppt
G9 Science Q4- Week 1-2 Projectile Motion.pptMAESTRELLAMesa2
 
Biopesticide (2).pptx .This slides helps to know the different types of biop...
Biopesticide (2).pptx  .This slides helps to know the different types of biop...Biopesticide (2).pptx  .This slides helps to know the different types of biop...
Biopesticide (2).pptx .This slides helps to know the different types of biop...RohitNehra6
 
Presentation Vikram Lander by Vedansh Gupta.pptx
Presentation Vikram Lander by Vedansh Gupta.pptxPresentation Vikram Lander by Vedansh Gupta.pptx
Presentation Vikram Lander by Vedansh Gupta.pptxgindu3009
 
Raman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral Analysis
Raman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral AnalysisRaman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral Analysis
Raman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral AnalysisDiwakar Mishra
 
Hubble Asteroid Hunter III. Physical properties of newly found asteroids
Hubble Asteroid Hunter III. Physical properties of newly found asteroidsHubble Asteroid Hunter III. Physical properties of newly found asteroids
Hubble Asteroid Hunter III. Physical properties of newly found asteroidsSérgio Sacani
 
Biological Classification BioHack (3).pdf
Biological Classification BioHack (3).pdfBiological Classification BioHack (3).pdf
Biological Classification BioHack (3).pdfmuntazimhurra
 
STERILITY TESTING OF PHARMACEUTICALS ppt by DR.C.P.PRINCE
STERILITY TESTING OF PHARMACEUTICALS ppt by DR.C.P.PRINCESTERILITY TESTING OF PHARMACEUTICALS ppt by DR.C.P.PRINCE
STERILITY TESTING OF PHARMACEUTICALS ppt by DR.C.P.PRINCEPRINCE C P
 
Lucknow 💋 Russian Call Girls Lucknow Finest Escorts Service 8923113531 Availa...
Lucknow 💋 Russian Call Girls Lucknow Finest Escorts Service 8923113531 Availa...Lucknow 💋 Russian Call Girls Lucknow Finest Escorts Service 8923113531 Availa...
Lucknow 💋 Russian Call Girls Lucknow Finest Escorts Service 8923113531 Availa...anilsa9823
 
Bentham & Hooker's Classification. along with the merits and demerits of the ...
Bentham & Hooker's Classification. along with the merits and demerits of the ...Bentham & Hooker's Classification. along with the merits and demerits of the ...
Bentham & Hooker's Classification. along with the merits and demerits of the ...Nistarini College, Purulia (W.B) India
 
SOLUBLE PATTERN RECOGNITION RECEPTORS.pptx
SOLUBLE PATTERN RECOGNITION RECEPTORS.pptxSOLUBLE PATTERN RECOGNITION RECEPTORS.pptx
SOLUBLE PATTERN RECOGNITION RECEPTORS.pptxkessiyaTpeter
 
PossibleEoarcheanRecordsoftheGeomagneticFieldPreservedintheIsuaSupracrustalBe...
PossibleEoarcheanRecordsoftheGeomagneticFieldPreservedintheIsuaSupracrustalBe...PossibleEoarcheanRecordsoftheGeomagneticFieldPreservedintheIsuaSupracrustalBe...
PossibleEoarcheanRecordsoftheGeomagneticFieldPreservedintheIsuaSupracrustalBe...Sérgio Sacani
 
Is RISC-V ready for HPC workload? Maybe?
Is RISC-V ready for HPC workload? Maybe?Is RISC-V ready for HPC workload? Maybe?
Is RISC-V ready for HPC workload? Maybe?Patrick Diehl
 
Natural Polymer Based Nanomaterials
Natural Polymer Based NanomaterialsNatural Polymer Based Nanomaterials
Natural Polymer Based NanomaterialsAArockiyaNisha
 
Animal Communication- Auditory and Visual.pptx
Animal Communication- Auditory and Visual.pptxAnimal Communication- Auditory and Visual.pptx
Animal Communication- Auditory and Visual.pptxUmerFayaz5
 
Nanoparticles synthesis and characterization​ ​
Nanoparticles synthesis and characterization​  ​Nanoparticles synthesis and characterization​  ​
Nanoparticles synthesis and characterization​ ​kaibalyasahoo82800
 
Cultivation of KODO MILLET . made by Ghanshyam pptx
Cultivation of KODO MILLET . made by Ghanshyam pptxCultivation of KODO MILLET . made by Ghanshyam pptx
Cultivation of KODO MILLET . made by Ghanshyam pptxpradhanghanshyam7136
 
Spermiogenesis or Spermateleosis or metamorphosis of spermatid
Spermiogenesis or Spermateleosis or metamorphosis of spermatidSpermiogenesis or Spermateleosis or metamorphosis of spermatid
Spermiogenesis or Spermateleosis or metamorphosis of spermatidSarthak Sekhar Mondal
 
Nightside clouds and disequilibrium chemistry on the hot Jupiter WASP-43b
Nightside clouds and disequilibrium chemistry on the hot Jupiter WASP-43bNightside clouds and disequilibrium chemistry on the hot Jupiter WASP-43b
Nightside clouds and disequilibrium chemistry on the hot Jupiter WASP-43bSérgio Sacani
 

Recently uploaded (20)

G9 Science Q4- Week 1-2 Projectile Motion.ppt
G9 Science Q4- Week 1-2 Projectile Motion.pptG9 Science Q4- Week 1-2 Projectile Motion.ppt
G9 Science Q4- Week 1-2 Projectile Motion.ppt
 
Biopesticide (2).pptx .This slides helps to know the different types of biop...
Biopesticide (2).pptx  .This slides helps to know the different types of biop...Biopesticide (2).pptx  .This slides helps to know the different types of biop...
Biopesticide (2).pptx .This slides helps to know the different types of biop...
 
Presentation Vikram Lander by Vedansh Gupta.pptx
Presentation Vikram Lander by Vedansh Gupta.pptxPresentation Vikram Lander by Vedansh Gupta.pptx
Presentation Vikram Lander by Vedansh Gupta.pptx
 
Raman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral Analysis
Raman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral AnalysisRaman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral Analysis
Raman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral Analysis
 
The Philosophy of Science
The Philosophy of ScienceThe Philosophy of Science
The Philosophy of Science
 
Hubble Asteroid Hunter III. Physical properties of newly found asteroids
Hubble Asteroid Hunter III. Physical properties of newly found asteroidsHubble Asteroid Hunter III. Physical properties of newly found asteroids
Hubble Asteroid Hunter III. Physical properties of newly found asteroids
 
Biological Classification BioHack (3).pdf
Biological Classification BioHack (3).pdfBiological Classification BioHack (3).pdf
Biological Classification BioHack (3).pdf
 
9953056974 Young Call Girls In Mahavir enclave Indian Quality Escort service
9953056974 Young Call Girls In Mahavir enclave Indian Quality Escort service9953056974 Young Call Girls In Mahavir enclave Indian Quality Escort service
9953056974 Young Call Girls In Mahavir enclave Indian Quality Escort service
 
STERILITY TESTING OF PHARMACEUTICALS ppt by DR.C.P.PRINCE
STERILITY TESTING OF PHARMACEUTICALS ppt by DR.C.P.PRINCESTERILITY TESTING OF PHARMACEUTICALS ppt by DR.C.P.PRINCE
STERILITY TESTING OF PHARMACEUTICALS ppt by DR.C.P.PRINCE
 
Lucknow 💋 Russian Call Girls Lucknow Finest Escorts Service 8923113531 Availa...
Lucknow 💋 Russian Call Girls Lucknow Finest Escorts Service 8923113531 Availa...Lucknow 💋 Russian Call Girls Lucknow Finest Escorts Service 8923113531 Availa...
Lucknow 💋 Russian Call Girls Lucknow Finest Escorts Service 8923113531 Availa...
 
Bentham & Hooker's Classification. along with the merits and demerits of the ...
Bentham & Hooker's Classification. along with the merits and demerits of the ...Bentham & Hooker's Classification. along with the merits and demerits of the ...
Bentham & Hooker's Classification. along with the merits and demerits of the ...
 
SOLUBLE PATTERN RECOGNITION RECEPTORS.pptx
SOLUBLE PATTERN RECOGNITION RECEPTORS.pptxSOLUBLE PATTERN RECOGNITION RECEPTORS.pptx
SOLUBLE PATTERN RECOGNITION RECEPTORS.pptx
 
PossibleEoarcheanRecordsoftheGeomagneticFieldPreservedintheIsuaSupracrustalBe...
PossibleEoarcheanRecordsoftheGeomagneticFieldPreservedintheIsuaSupracrustalBe...PossibleEoarcheanRecordsoftheGeomagneticFieldPreservedintheIsuaSupracrustalBe...
PossibleEoarcheanRecordsoftheGeomagneticFieldPreservedintheIsuaSupracrustalBe...
 
Is RISC-V ready for HPC workload? Maybe?
Is RISC-V ready for HPC workload? Maybe?Is RISC-V ready for HPC workload? Maybe?
Is RISC-V ready for HPC workload? Maybe?
 
Natural Polymer Based Nanomaterials
Natural Polymer Based NanomaterialsNatural Polymer Based Nanomaterials
Natural Polymer Based Nanomaterials
 
Animal Communication- Auditory and Visual.pptx
Animal Communication- Auditory and Visual.pptxAnimal Communication- Auditory and Visual.pptx
Animal Communication- Auditory and Visual.pptx
 
Nanoparticles synthesis and characterization​ ​
Nanoparticles synthesis and characterization​  ​Nanoparticles synthesis and characterization​  ​
Nanoparticles synthesis and characterization​ ​
 
Cultivation of KODO MILLET . made by Ghanshyam pptx
Cultivation of KODO MILLET . made by Ghanshyam pptxCultivation of KODO MILLET . made by Ghanshyam pptx
Cultivation of KODO MILLET . made by Ghanshyam pptx
 
Spermiogenesis or Spermateleosis or metamorphosis of spermatid
Spermiogenesis or Spermateleosis or metamorphosis of spermatidSpermiogenesis or Spermateleosis or metamorphosis of spermatid
Spermiogenesis or Spermateleosis or metamorphosis of spermatid
 
Nightside clouds and disequilibrium chemistry on the hot Jupiter WASP-43b
Nightside clouds and disequilibrium chemistry on the hot Jupiter WASP-43bNightside clouds and disequilibrium chemistry on the hot Jupiter WASP-43b
Nightside clouds and disequilibrium chemistry on the hot Jupiter WASP-43b
 

Unbiased Markov chain Monte Carlo methods

  • 1. Unbiased Markov chain Monte Carlo methods Pierre E. Jacob Department of Statistics, Harvard University joint work with John O’Leary, Yves Atchad´e, and other fantastic people acknowledged throughout Department of Statistics, Stanford University April 16, 2019 Pierre E. Jacob Unbiased MCMC
  • 2. Outline 1 Monte Carlo and bias 2 Unbiased estimators from Markov kernels 3 Numerical experiments 4 Beyond parallel computation Pierre E. Jacob Unbiased MCMC
  • 3. Outline 1 Monte Carlo and bias 2 Unbiased estimators from Markov kernels 3 Numerical experiments 4 Beyond parallel computation Pierre E. Jacob Unbiased MCMC
  • 4. Setting Continuous or discrete space of dimension d. Target probability distribution π, with probability density function x → π(x). Goal: approximate π, e.g. approximate Eπ[h(X)] = h(x)π(x)dx = π(h), for a class of “test” functions h. Pierre E. Jacob Unbiased MCMC
  • 5. Markov chain Monte Carlo Initially, X0 ∼ π0, then Xt|Xt−1 ∼ P(Xt−1, ·) for t = 1, . . . , T. Estimator: 1 T − b T t=b+1 h(Xt), where b iterations are discarded as burn-in. Might converge to Eπ[h(X)] as T → ∞ by the ergodic theorem. Biased for any fixed b, T, since π0 = π. Averaging independent copies of such estimators for fixed T, b would not provide a consistent estimator of Eπ[h(X)] as the number of independent copies goes to infinity. Pierre E. Jacob Unbiased MCMC
  • 6. Example: Metropolis–Hastings kernel P Initialize: X0 ∼ π0. At each iteration t ≥ 0, with Markov chain at state Xt, 1 propose X ∼ k(Xt, ·), 2 sample U ∼ U([0, 1]), 3 if U ≤ π(X )k(X , Xt) π(Xt)k(Xt, X ) , set Xt+1 = X , otherwise set Xt+1 = Xt. Hastings, Monte Carlo sampling methods using Markov chains and their applications, Biometrika, 1970. Pierre E. Jacob Unbiased MCMC
  • 7. MCMC path π = N(0, 1), MH with Normal proposal std = 0.5, π0 = N(10, 32 ) Pierre E. Jacob Unbiased MCMC
  • 8. MCMC marginal distributions π = N(0, 1), RWMH with Normal proposal std = 0.5, π0 = N(10, 32 ) Pierre E. Jacob Unbiased MCMC
  • 9. Sequential Monte Carlo samplers π = N(0, 1), adaptive SMC sampler with MH moves, π0 = N(10, 32 ) Neal, Annealed importance sampling, STCO, 2001. Del Moral, Doucet, Jasra, SMC samplers, JRSS B, 2006. Pierre E. Jacob Unbiased MCMC
  • 10. Empirical measure approximations MCMC: ˆπ = 1 T T t=1 δXt → π as T → ∞. SMC: ˆπ = N n=1 ωnδXn → π as N → ∞. Proposed: ˆπ = N n=1 ωnδXn , with E[Eˆπ[h(X)]] = Eπ[h(X)]. Pierre E. Jacob Unbiased MCMC
  • 11. Do we care about unbiased estimators? In classical point estimation, unbiasedness is not crucial. Larry Wasserman in “All of Statistics” (2003) writes: Unbiasedness used to receive much attention but these days is considered less important. On the other hand, Jeff Rosenthal in “Parallel computing and Monte Carlo algorithms” (2000) writes: When running parallel Monte Carlo with many computers, it is more important to start with an unbiased (or low-bias) estimate than with a low-variance estimate. Pierre E. Jacob Unbiased MCMC
  • 12. Parallel computing R processors generate unbiased estimators, independently. The number of estimators completed by time t is random. There are different ways of aggregating estimators, either discarding on-going calculations at time t or not. Different regimes can be considered, e.g. as R → ∞ for fixed t, as t → ∞ for fixed R, as t and R both go to infinity. Glynn & Heidelberger, Analysis of parallel replicated simulations under a completion time constraint, 1991. Pierre E. Jacob Unbiased MCMC
  • 13. Parallel computing Pierre E. Jacob Unbiased MCMC
  • 14. Outline 1 Monte Carlo and bias 2 Unbiased estimators from Markov kernels 3 Numerical experiments 4 Beyond parallel computation Pierre E. Jacob Unbiased MCMC
  • 15. Coupled chains Generate two Markov chains (Xt) and (Yt) as follows, sample X0 and Y0 from π0 (independently or not), sample X1|X0 ∼ P(X0, ·), for t ≥ 1, sample (Xt+1, Yt)|(Xt, Yt−1) ∼ ¯P ((Xt, Yt−1), ·), ¯P is such that Xt+1|Xt ∼ P(Xt, ·) and Yt|Yt−1 ∼ P(Yt−1, ·), so each chain marginally evolves according to P. ⇒ Xt and Yt have the same distribution for all t ≥ 0. ¯P is also such that ∃τ such that Xτ = Yτ−1, and the chains are faithful. Pierre E. Jacob Unbiased MCMC
  • 16. Debiasing idea (one slide version) Limit as a telescopic sum, for all k ≥ 0, Eπ[h(X)] = lim t→∞ E[h(Xt)] = E[h(Xk)] + ∞ t=k+1 E[h(Xt) − h(Xt−1)]. Since for all t ≥ 0, Xt and Yt have the same distribution, = E[h(Xk)] + ∞ t=k+1 E[h(Xt) − h(Yt−1)]. If we can swap expectation and limit, = E[h(Xk) + ∞ t=k+1 (h(Xt) − h(Yt−1))], so h(Xk) + ∞ t=k+1(h(Xt) − h(Yt−1)) is unbiased. Pierre E. Jacob Unbiased MCMC
  • 17. Unbiased estimators Unbiased estimator is given by Hk(X, Y ) = h(Xk) + τ−1 t=k+1 (h(Xt) − h(Yt−1)), with the convention τ−1 t=k+1{·} = 0 if τ − 1 < k + 1. h(Xk) alone is biased; the other terms correct for the bias. Cost: τ − 1 calls to ¯P and 1 + max(0, k − τ) calls to P. Glynn & Rhee, Exact estimation for Markov chain equilibrium expectations, 2014. Pierre E. Jacob Unbiased MCMC
  • 18. Conditions Jacob, O’Leary, Atchad´e, Unbiased MCMC with couplings. 1 Marginal chain converges: E[h(Xt)] → Eπ[h(X)], and h(Xt) has (2 + η)-finite moments for all t. 2 Meeting time τ has geometric tails: ∃C < +∞ ∃δ ∈ (0, 1) ∀t ≥ 0 P(τ > t) ≤ Cδt . 3 Chains stay together: Xt = Yt−1 for all t ≥ τ. Condition 2 itself implied by e.g. geometric drift condition. Under these conditions, Hk(X, Y ) is unbiased, has finite expected cost and finite variance, for all k. Pierre E. Jacob Unbiased MCMC
  • 19. Conditions: update Middleton, Deligiannidis, Doucet, Jacob, Unbiased MCMC for intractable target distributions. 1 Marginal chain converges: E[h(Xt)] → Eπ[h(X)], and h(Xt) has (2 + η)-finite moments for all t. 2 Meeting time τ has polynomial tails: ∃C < +∞ ∃δ > 2(2η−1 + 1) ∀t ≥ 0 P(τ > t) ≤ Ct−δ . 3 Chains stay together: Xt = Yt−1 for all t ≥ τ. Condition 2 itself implied by e.g. polynomial drift condition. Pierre E. Jacob Unbiased MCMC
  • 20. Improved unbiased estimators Efficiency matters, thus in practice we recommend a variation of the previous estimator, defined for integers k ≤ m as Hk:m(X, Y ) = 1 m − k + 1 m t=k Ht(X, Y ) which can also be written 1 m − k + 1 m t=k h(Xt)+ τ−1 t=k+1 min 1, t − k m − k + 1 (h(Xt)−h(Yt−1)), i.e. standard MCMC average + bias correction term. As k → ∞, bias correction is zero with increasing probability. Pierre E. Jacob Unbiased MCMC
  • 21. Efficiency Writing Hk:m(X, Y ) = MCMCk:m + BCk:m, then, denoting the MSE of MCMCk:m by MSEk:m, V[Hk:m(X, Y )] ≤ MSEk:m+2 MSEk:m E BC2 k:m +E BC2 k:m . Under geometric drift condition and regularity assumptions on h, for some δ < 1, C < ∞, E[BC2 k:m] ≤ Cδk (m − k + 1)2 , and δ is directly related to tails of the meeting time. Similarly under polynomial drift, we obtain a term of the form k−δ in the numerator instead of δk. Pierre E. Jacob Unbiased MCMC
  • 22. Signed measure estimator Replacing function evaluations by delta masses leads to ˆπ(·) = 1 m − k + 1 m t=k δXt (·) + τ−1 t=k+1 min 1, t − k m − k + 1 (δXt (·) − δYt−1 (·)), which is of the form ˆπ(·) = N n=1 ωnδXn (·), where N n=1 ωn = 1 but some ωn might be negative. Unbiasedness reads: E[Eˆπ[h(X)]] = Eπ[h(X)]. Pierre E. Jacob Unbiased MCMC
  • 23. Determining a warm-up period Total variation distance between Xk ∼ πk and π = limk→∞ πk: πk − π TV = 1 2 sup h:|h|≤1 |E[h(Xk)] − Eπ[h(X)]| = 1 2 sup h:|h|≤1 |E[ τ−1 t=k+1 h(Xt) − h(Yt−1)]| ≤ E[max(0, τ − k + 1)]. 0.00 0.01 0.02 0.03 0.04 0 50 100 150 200 meeting time density 1e−03 1e−02 1e−01 1e+00 1e+01 0 50 100 k upperbound Pierre E. Jacob Unbiased MCMC
  • 24. Designing coupled chains To implement the proposed unbiased estimators, we need to sample from a Markov kernel ¯P, such that, when (Xt+1, Yt) is sampled from ¯P ((Xt, Yt−1), ·), marginally Xt+1|Xt ∼ P(Xt, ·), and Yt|Yt−1 ∼ P(Yt−1, ·), it is possible that Xt+1 = Yt exactly for some t ≥ 0, if Xt = Yt−1, then Xt+1 = Yt almost surely. Pierre E. Jacob Unbiased MCMC
  • 25. Couplings of MCMC algorithms Many practical couplings in the literature. . . Propp & Wilson, Exact sampling with coupled Markov chains and applications to statistical mechanics, Random Structures & Algorithms, 1996. Johnson, Studying convergence of Markov chain Monte Carlo algorithms using coupled sample paths, JASA, 1996. Neal, Circularly-coupled Markov chain sampling, UoT tech report, 1999. Pinto & Neal, Improving Markov chain Monte Carlo estimators by coupling to an approximating chain, UoT tech report, 2001. Glynn & Rhee, Exact estimation for Markov chain equilibrium expectations, Journal of Applied Probability, 2014. Pierre E. Jacob Unbiased MCMC
  • 26. Couplings of MCMC algorithms: contributions Conditional particle filters Jacob, Lindsten, Sch¨on, Smoothing with Couplings of Conditional Particle Filters, JASA, 2018. Metropolis–Hastings, Gibbs samplers, parallel tempering Jacob, O’Leary, Atchad´e, Unbiased MCMC with couplings, submitted. Hamiltonian Monte Carlo Heng & Jacob, Unbiased HMC with couplings, Biometrika, 2019. Pseudo-marginal MCMC, exchange algorithm Middleton, Deligiannidis, Doucet, Jacob, Unbiased MCMC for intractable target distributions, submitted. Particle independent Metropolis–Hastings Middleton, Deligiannidis, Doucet, Jacob, Unbiased Smoothing using Particle Independent Metropolis-Hastings, AISTATS, 2019. Pierre E. Jacob Unbiased MCMC
  • 27. Couplings (X, Y ) follows a coupling of p and q if X ∼ p and Y ∼ q. The coupling inequality states that P(X = Y ) ≤ 1 − p − q TV, for any coupling, with p − q TV = 1 2 |p(x) − q(x)|dx. Maximal couplings achieve the bound. Pierre E. Jacob Unbiased MCMC
  • 28. Maximal coupling of Gamma and Normal Pierre E. Jacob Unbiased MCMC
  • 29. Maximal coupling: sampling algorithm Requires: evaluations of p and q, sampling from p and q. 1 Sample X ∼ p and W ∼ U([0, 1]). If W ≤ q(X)/p(X), output (X, X). 2 Otherwise, sample Y ∼ q and W ∼ U([0, 1]) until W > p(Y )/q(Y ), and output (X, Y ). Output: a pair (X, Y ) such that X ∼ p, Y ∼ q and P(X = Y ) is maximal. Pierre E. Jacob Unbiased MCMC
  • 30. Back to Metropolis–Hastings (kernel P) At each iteration t, Markov chain at state Xt, 1 propose X ∼ k(Xt, ·), 2 sample U ∼ U([0, 1]), 3 if U ≤ π(X )k(X , Xt) π(Xt)k(Xt, X ) , set Xt+1 = X , otherwise set Xt+1 = Xt. How to propagate two MH chains from states Xt and Yt−1 such that {Xt+1 = Yt} can happen? Pierre E. Jacob Unbiased MCMC
  • 31. Coupling of Metropolis–Hastings (kernel ¯P) At each iteration t, two Markov chains at states Xt, Yt−1, 1 propose (X , Y ) from max coupling of k(Xt, ·), k(Yt−1, ·), 2 sample U ∼ U([0, 1]), 3 if U ≤ π(X )k(X , Xt) π(Xt)k(Xt, X ) , set Xt+1 = X , otherwise set Xt+1 = Xt, if U ≤ π(Y )k(Y , Yt−1) π(Yt−1)k(Yt−1, Y ) , set Yt = Y , otherwise set Yt = Yt−1. Pierre E. Jacob Unbiased MCMC
  • 32. Coupling of Gibbs Gibbs sampler: update component i of the chain, leaving π(dxi|x1, . . . , xi−1, xi+1, . . . , xd) invariant. For instance, we can propose X ∼ k(Xi t, ·) to replace Xi t, and accept or not with a Metropolis–Hastings step. These proposals can be maximally coupled across two chains, at each component update. The chains meet when all components have met. Likewise we can couple parallel tempering chains, and meeting occurs when entire ensembles of chains meet. Pierre E. Jacob Unbiased MCMC
  • 33. Hamiltonian Monte Carlo Introduce potential energy U(q) = − log π(q), and total energy E(q, p) = U(q) + 1 2|p|2. Hamiltonian dynamics for (q(s), p(s)), where s ≥ 0: d ds q(s) = pE(q(s), p(s)), d ds p(s) = − qE(q(s), p(s)). Solving Hamiltonian dynamics exactly is not feasible, but discretization + Metropolis–Hastings correction ensure that π remains invariant. Common random numbers can make two HMC chains contract, under assumptions on the target such as strong log-concavity. Heng & Jacob, Unbiased HMC with couplings, Biometrika, 2019. Pierre E. Jacob Unbiased MCMC
  • 34. Coupling of HMC Figure 2 of Mangoubi & Smith, Rapid mixing of HMC strongly log-concave distributions, 2017. Coupling two copies X1, X2, . . . (blue) and Y1, Y2, . . . (green) of HMC by choosing same momentum pi at every step. Pierre E. Jacob Unbiased MCMC
  • 35. Particle Independent Metropolis–Hastings Initialization: run SMC, obtain ˆπ(0)(·), ˆZ(0), where E[ ˆZ(0)] = Z, the normalizing constant of π. At each iteration t, given (ˆπ(t)(·), ˆZ(t)), 1 run SMC, obtain (π (·), Z ), 2 sample U ∼ U([0, 1]), 3 if U ≤ Z ˆZ(t) , set (t + 1)-th state to (π (·), Z ), otherwise set (t + 1)-th state to (ˆπ(t)(·), ˆZ(t)). For any number of particles, T−1 T t=1 ˆπ(t)(·) →T→∞ π. Andrieu, Doucet, Holenstein, Particle MCMC, JRSS B, 2010. Pierre E. Jacob Unbiased MCMC
  • 36. Coupled Particle Independent Metropolis–Hastings At each iteration t, given (ˆπ(t)(·), ˆZ(t)), and (˜π(t−1)(·), ˜Z(t−1)), 1 run SMC, obtain (π (·), Z ), 2 sample U ∼ U([0, 1]), 3 if U ≤ Z ˆZ(t) , set (t + 1)-th “1st” state to (π (·), Z ), otherwise set (t + 1)-th “1st” state to (ˆπ(t)(·), ˆZ(t)). 4 if U ≤ Z ˜Z(t−1) , set t-th “2nd” state to (π (·), Z ), otherwise set t-th “2nd” state to (˜π(t−1)(·), ˜Z(t−1)). Pierre E. Jacob Unbiased MCMC
  • 37. Unbiased SMC PIMH turns any SMC sampler into an MCMC scheme that can be debiased using the coupled chains machinery. Thus any MCMC algorithm that can be plugged in an SMC sampler, can then be subsequently de-biased. No need for couplings of the MCMC chains themselves. Middleton, Deligiannidis, Doucet, Jacob, Unbiased Smoothing using Particle Independent Metropolis-Hastings, AISTATS, 2019. Pierre E. Jacob Unbiased MCMC
  • 38. Outline 1 Monte Carlo and bias 2 Unbiased estimators from Markov kernels 3 Numerical experiments 4 Beyond parallel computation Pierre E. Jacob Unbiased MCMC
  • 39. Linear regression: model Johndrow, Orenstein, Bhattacharya, Bayes Shrinkage at GWAS scale: Convergence and Approximation Theory of a Scalable MCMC Algorithm for the Horseshoe Prior. ∀i ∈ {1, . . . , n} Yi|Xi, β, σ2 ∼ N(XT i β, σ2 ) ∀j ∈ {1, . . . , p} βj|σ2 , ηj, ξ ∼ N(0, σ2 ξ η−1 j ), η −1/2 j ∼ C+ (0, 1) ξ−1/2 ∼ C+ (0, 1), σ2 ∼ InvGamma( 1 2 , 1 2 ) Data generated with Xij i.i.d. ∼ N(0, 1), σ2 = 0.12, βj = 4 for j = 1, . . . , 5, βj = 2(6−j)/2 for j = 6, . . . , 20, βj = 0 for j ≥ 21. Pierre E. Jacob Unbiased MCMC
  • 40. Linear regression: Gibbs sampler η|β, ξ, σ2 from distribution ∝ p j=1 1 1+ηj exp(− β2 j ξ 2σ2 ηj)1(ηj > 0), using rejection sampler for each component j, ⇒ maximal coupling for each component j, ξ|η using Metropolis–Hastings step, ⇒ maximal coupling of the Normal proposals, σ2|η, ξ from Inverse Gamma distribution, ⇒ maximal coupling, β|η, ξ, σ2 from p-dimensional Normal distribution, ⇒ common random numbers in fast algorithm from Bhattacharya, Chakraborty, Mallick (2016). Pierre E. Jacob Unbiased MCMC
  • 41. Linear regression: meeting times p=100 p=250 p=500 n=100n=250n=500 10 100 1000 10000 10 100 1000 10000 10 100 1000 10000 0 5 10 15 20 0 5 10 15 20 0 5 10 15 20 meeting time count Biswas & Jacob, Unbiased MCMC for linear regressions. Pierre E. Jacob Unbiased MCMC
  • 42. Linear regression: meeting times p=100 p=250 p=500 n=100n=250n=500 1 10 100 100010000 1 10 100 100010000 1 10 100 100010000 0 10 20 0 10 20 0 10 20 duration (seconds) count Biswas & Jacob, Unbiased MCMC for linear regressions. Pierre E. Jacob Unbiased MCMC
  • 43. Whisker experiment Activation of neurons of rats, as their whiskers are being moved with a periodic stimulus. The activation of a neuron is recorded as a binary variable for each time and each experiment. These activation variables are then aggregated by summing over the M experiments at each time step. Data kindly shared by Demba Ba (SEAS). Lin, Zhang, Heng, Allsop, Tye, Jacob, Ba, Clustering Time Series with Nonlinear Dynamics: A Bayesian Non-Parametric and Particle-Based Approach, AISTATS 2019. Pierre E. Jacob Unbiased MCMC
  • 44. Whisker experiment: data Model: X0 ∼ N(0, 1), Xt|Xt−1 ∼ N(aXt−1, σ2 X), Yt|Xt ∼ Binom(M, (1 + exp(−Xt))−1 ). Pierre E. Jacob Unbiased MCMC
  • 45. Whisker experiment: meeting times Middleton, Deligiannidis, Doucet, Jacob, Unbiased MCMC for intractable target distributions. Heng, Bishop, Deligiannidis & Doucet, Controlled SMC. Pierre E. Jacob Unbiased MCMC
  • 46. Whisker experiment: in parallel Computation chronology, on 100 processors over 23 hours, with k = 1, 000, m = 10k. Pierre E. Jacob Unbiased MCMC
  • 47. Whisker experiment: results Approximations of marginal posteriors, k = 1, 000, m = 10k. Red line corresponds to PMMH run (250,000 iterations). Pierre E. Jacob Unbiased MCMC
  • 48. Outline 1 Monte Carlo and bias 2 Unbiased estimators from Markov kernels 3 Numerical experiments 4 Beyond parallel computation Pierre E. Jacob Unbiased MCMC
  • 49. Unbiased property Unbiased MCMC estimators can be used to tackle problems for which standard MCMC methods are ill-suited. For instance, problems where one needs to approximate h(x1, x2)π2(x2|x1)dx2 π1(x1)dx1, or equivalently E1 [E2[h(X1, X2)|X1]] . Pierre E. Jacob Unbiased MCMC
  • 50. Example 1: cut distribution First model, parameter θ1, data Y1, posterior π1(θ1|Y1). Second model, parameter θ2, data Y2, conditional posterior: π2(θ2|Y2, θ1) = p2(θ2|θ1)p2(Y2|θ1, θ2) p2(Y2|θ1) . Plummer, Cuts in Bayesian graphical models, 2015. Jacob, Murray, Holmes, Robert, Better together? Statistical learning in models made of modules. Pierre E. Jacob Unbiased MCMC
  • 51. Example 2: normalizing constants We might be interested in normalizing constant Z of π(x) ∝ exp(−U(x)), defined as Z = exp(−U(x))dx. Introduce ∀λ ∈ [0, 1] πλ(x) ∝ exp(−Uλ(x)). Thermodynamics integration or path sampling identity: log Z1 Z0 = − 1 0 λUλ(x)πλ(dx) q(λ) q(dλ), where q(dλ) is any distribution supported on [0, 1]. Rischard, Jacob, Pillai, Unbiased estimation of log normalizing constants with applications to Bayesian cross-validation. Pierre E. Jacob Unbiased MCMC
  • 52. Example 3: Bayesian cross-validation Data y = {y1, . . . , yn}, made of n units. Partition y into training T and validation V sets. For each split y = {T, V }, assess predictive performance with log p(V |T) = log p(V |θ, T)p(dθ|T). Note, log p(V |T) is a log-ratio of normalizing constants. Finally we might want to approximate CV = n nT −1 T,V log p(V |T). Rischard, Jacob, Pillai, Unbiased estimation of log normalizing constants with applications to Bayesian cross-validation. Pierre E. Jacob Unbiased MCMC
  • 53. Discussion Perfect samplers, designed to sample i.i.d. from π, would yield the same benefits and more. What about regeneration? Implications of lack of bias are numerous. Proposed estimators can be obtained using coupled MCMC kernels, or using the “MCMC → SMC → coupled PIMH” route. Cannot work if underlying MCMC doesn’t work. . . but parallelism allows for more expensive MCMC kernels. Thank you for listening, and special thanks to James Johndrow! Pierre E. Jacob Unbiased MCMC
  • 54. References with Fredrik Lindsten, Thomas Sch¨on Smoothing with Couplings of Conditional Particle Filters, 2018. with John O’Leary, Yves F. Atchad´e Unbiased Markov chain Monte Carlo with couplings, 2019. with Jeremy Heng Unbiased Hamiltonian Monte Carlo with couplings, 2019. with Lawrence Middleton, George Deligiannidis, Arnaud Doucet Unbiased Markov chain Monte Carlo for intractable target distributions, 2019. Unbiased Smoothing using Particle Independent Metropolis-Hastings, 2019. with Maxime Rischard, Natesh Pillai Unbiased estimation of log normalizing constants with applications to Bayesian cross-validation, 2019. Pierre E. Jacob Unbiased MCMC