## NIPS 2014

Posted in Kids, pictures, Statistics, Travel, University life with tags , , , , , , , , , , , on December 15, 2014 by xi'an

Second and last day of the NIPS workshops! The collection of topics was quite broad and would have made my choosing an ordeal, except that I was invited to give a talk at the probabilistic programming workshop, solving my dilemma… The first talk by Kathleen Fisher was quite enjoyable in that it gave a conceptual discussion of the motivations for probabilistic languages, drawing an analogy with the early days of computer programming that saw a separation between higher level computer languages and machine programming, with a compiler interface. And calling for a similar separation between the models faced by statistical inference and machine-learning and the corresponding code, if I understood her correctly. This was connected with Frank Wood’s talk of the previous day where he illustrated the concept through a generation of computer codes to approximately generate from standard distributions like Normal or Poisson. Approximately as in ABC, which is why the organisers invited me to talk in this session. However, I was a wee bit lost in the following talks and presumably lost part of my audience during my talk, as I realised later to my dismay when someone told me he had not perceived the distinction between the trees in the random forest procedure and the phylogenetic trees in the population genetic application. Still, while it had for me a sort of Twilight Zone feeling of having stepped in another dimension, attending this workshop was an worthwhile experiment as an eye-opener into a highly different albeit connected field, where code and simulator may take the place of a likelihood function… To the point of defining Hamiltonian Monte Carlo directly on the former, as Vikash Mansinghka showed me at the break.

I completed the day with the final talks in the variational inference workshop, if only to get back on firmer ground! Apart from attending my third talk by Vikash in the conference (but on a completely different topic on variational approximations for discrete particle-ar distributions), a talk by Tim Salimans linked MCMC and variational approximations, using MCMC and HMC to derive variational bounds. (He did not expand on the opposite use of variational approximations to build better proposals.) Overall, I found these two days and my first NIPS conference quite exciting, if somewhat overpowering, with a different atmosphere and a different pace compared with (small or large) statistical meetings. (And a staggering gender imbalance!)

## ABC à Montréal

Posted in Kids, pictures, Running, Statistics, Travel, University life with tags , , , , , , , , , , , , , , , , on December 13, 2014 by xi'an

So today was the NIPS 2014 workshop, “ABC in Montréal“, which started with a fantastic talk by Juliane Liepe on some exciting applications of ABC to the migration of immune cells, with the analysis of movies involving those cells acting to heal a damaged fly wing and a cut fish tail. Quite amazing videos, really. (With the great entry line of ‘We have all cut  a finger at some point in our lives’!) The statistical model behind those movies was a random walk on a grid, with different drift and bias features that served as model characteristics. Frank Wood managed to deliver his talk despite a severe case of food poisoning, with a great illustration of probabilistic programming that made me understand (at last!) the very idea of probabilistic programming. And  Vikash Mansinghka presented some applications in image analysis. Those two talks led me to realise why probabilistic programming was so close to ABC, with a programming touch! Hence why I was invited to talk today! Then Dennis Prangle exposed his latest version of lazy ABC, that I have already commented on the ‘Og, somewhat connected with our delayed acceptance algorithm, to the point that maybe something common can stem out of the two notions. Michael Blum ended the day with provocative answers to the provocative question of Ted Meeds as to whether or not machine learning needed ABC (Ans. No!) and whether or not machine learning could help ABC (Ans. ???). With an happily mix-up between mechanistic and phenomenological models that helped generating discussion from the floor.

The posters were also of much interest, with calibration as a distance measure by Michael Guttman, in continuation of the poster he gave at MCMski, Aaron Smith presenting his work with Luke Bornn, Natesh Pillai and Dawn Woodard, on why a single pseudo-sample is enough for ABC efficiency. This gave me the opportunity to discuss with him the apparent contradiction with the result of Kryz Łatunsziński and Anthony Lee about the geometric convergence of ABC-MCMC only attained with a random number of pseudo-samples… And to wonder if there is a geometric versus binomial dilemma in this setting, Namely, whether or not simulating pseudo-samples until one is accepted would be more efficient than just running one and discarding it in case it is too far. So, although the audience was not that large (when compared with the other “ABC in…” and when considering the 2500+ attendees at NIPS over the week!), it was a great day where I learned a lot, did not have a doze during talks (!), [and even had an epiphany of sorts at the treadmill when I realised I just had to take longer steps to reach 16km/h without hyperventilating!] So thanks to my fellow organisers, Neil D Lawrence, Ted Meeds, Max Welling, and Richard Wilkinson for setting the program of that day! And, by the way, where’s the next “ABC in…”?! (Finland, maybe?)

## reflections on the probability space induced by moment conditions with implications for Bayesian Inference [slides]

Posted in Books, Statistics, University life with tags , , , , , , , , , , , , on December 4, 2014 by xi'an

Here are the slides of my incoming discussion of Ron Gallant’s paper, tomorrow.

## another instance of ABC?

Posted in Statistics with tags , , , , , on December 2, 2014 by xi'an

“These characteristics are (1) likelihood is not available; (2) prior information is available; (3) a portion of the prior information is expressed in terms of functionals of the model that cannot be converted into an analytic prior on model parameters; (4) the model can be simulated. Our approach depends on an assumption that (5) an adequate statistical model for the data are available.”

A 2009 JASA paper by Ron Gallant and Rob McCulloch, entitled “On the Determination of General Scientific Models With Application to Asset Pricing”, may have or may not have connection with ABC, to wit the above quote, but I have trouble checking whether or not this is the case.

The true (scientific) model parametrised by θ is replaced with a (statistical) substitute that is available in closed form. And parametrised by g(θ). [If you can get access to the paper, I’d welcome opinions about Assumption 1 therein which states that the intractable density is equal to a closed-form density.] And the latter is over-parametrised when compared with the scientific model. As in, e.g., a N(θ,θ²) scientific model versus a N(μ,σ²) statistical model. In addition, the prior information is only available on θ. However, this does not seem to matter that much since (a) the Bayesian analysis is operated on θ only and (b) the Metropolis approach adopted by the authors involves simulating a massive number of pseudo-observations, given the current value of the parameter θ and the scientific model, so that the transform g(θ) can be estimated by maximum likelihood over the statistical model. The paper suggests using a secondary Markov chain algorithm to find this MLE. Which is claimed to be a simulated annealing resolution (p.121) although I do not see the temperature decreasing. The pseudo-model is then used in a primary MCMC step.

Hence, not truly an ABC algorithm. In the same setting, ABC would use a simulated dataset the same size as the observed dataset, compute the MLEs for both and compare them. Faster if less accurate when Assumption 1 [that the statistical model holds for a restricted parametrisation] does not stand.

Another interesting aspect of the paper is about creating and using a prior distribution around the manifold η=g(θ). This clearly relates to my earlier query about simulating on measure zero sets. The paper does not bring a definitive answer, as it never simulates exactly on the manifold, but this constitutes another entry on this challenging problem…

## reflections on the probability space induced by moment conditions with implications for Bayesian Inference [refleXions]

Posted in Statistics, University life with tags , , , , , , , , , , on November 26, 2014 by xi'an

“The main finding is that if the moment functions have one of the properties of a pivotal, then the assertion of a distribution on moment functions coupled with a proper prior does permit Bayesian inference. Without the semi-pivotal condition, the assertion of a distribution for moment functions either partially or completely specifies the prior.” (p.1)

Ron Gallant will present this paper at the Conference in honour of Christian Gouréroux held next week at Dauphine and I have been asked to discuss it. What follows is a collection of notes I made while reading the paper , rather than a coherent discussion, to come later. Hopefully prior to the conference.

The difficulty I have with the approach presented therein stands as much with the presentation as with the contents. I find it difficult to grasp the assumptions behind the model(s) and the motivations for only considering a moment and its distribution. Does it all come down to linking fiducial distributions with Bayesian approaches? In which case I am as usual sceptical about the ability to impose an arbitrary distribution on an arbitrary transform of the pair (x,θ), where x denotes the data. Rather than a genuine prior x likelihood construct. But I bet this is mostly linked with my lack of understanding of the notion of structural models.

“We are concerned with situations where the structural model does not imply exogeneity of θ, or one prefers not to rely on an assumption of exogeneity, or one cannot construct a likelihood at all due to the complexity of the model, or one does not trust the numerical approximations needed to construct a likelihood.” (p.4)

As often with econometrics papers, this notion of structural model sets me astray: does this mean any latent variable model or an incompletely defined model, and if so why is it incompletely defined? From a frequentist perspective anything random is not a parameter. The term exogeneity also hints at this notion of the parameter being not truly a parameter, but including latent variables and maybe random effects. Reading further (p.7) drives me to understand the structural model as defined by a moment condition, in the sense that

$\mathbb{E}[m(\mathbf{x},\theta)]=0$

has a unique solution in θ under the true model. However the focus then seems to make a major switch as Gallant considers the distribution of a pivotal quantity like

$Z=\sqrt{n} W(\mathbf{x},\theta)^{-\frac{1}{2}} m(\mathbf{x},\theta)$

as induced by the joint distribution on (x,θ), hence conversely inducing constraints on this joint, as well as an associated conditional. Which is something I have trouble understanding, First, where does this assumed distribution on Z stem from? And, second, exchanging randomness of terms in a random variable as if it was a linear equation is a pretty sure way to produce paradoxes and measure theoretic difficulties.

The purely mathematical problem itself is puzzling: if one knows the distribution of the transform Z=Z(X,Λ), what does that imply on the joint distribution of (X,Λ)? It seems unlikely this will induce a single prior and/or a single likelihood… It is actually more probable that the distribution one arbitrarily selects on m(x,θ) is incompatible with a joint on (x,θ), isn’t it?

“The usual computational method is MCMC (Markov chain Monte Carlo) for which the best known reference in econometrics is Chernozhukov and Hong (2003).” (p.6)

While I never heard of this reference before, it looks like a 50 page survey and may be sufficient for an introduction to MCMC methods for econometricians. What I do not get though is the connection between this reference to MCMC and the overall discussion of constructing priors (or not) out of fiducial distributions. The author also suggests using MCMC to produce the MAP estimate but this always stroke me as inefficient (unless one uses our SAME algorithm of course).

“One can also compute the marginal likelihood from the chain (Newton and Raftery (1994)), which is used for Bayesian model comparison.” (p.22)

Not the best solution to rely on harmonic means for marginal likelihoods…. Definitely not. While the author actually uses the stabilised version (15) of Newton and Raftery (1994) estimator, which in retrospect looks much like a bridge sampling estimator of sorts, it remains dangerously close to the original [harmonic mean solution] especially for a vague prior. And it only works when the likelihood is available in closed form.

“The MCMC chains were comprised of 100,000 draws well past the point where transients died off.” (p.22)

I wonder if the second statement (with a very nice image of those dying transients!) is intended as a consequence of the first one or independently.

“A common situation that requires consideration of the notions that follow is that deriving the likelihood from a structural model is analytically intractable and one cannot verify that the numerical approximations one would have to make to circumvent the intractability are sufficiently accurate.” (p.7)

This then is a completely different business, namely that defining a joint distribution by mean of moment equations prevents regular Bayesian inference because the likelihood is not available. This is more exciting because (i) there are alternative available! From ABC to INLA (maybe) to EP to variational Bayes (maybe). And beyond. In particular, the moment equations are strongly and even insistently suggesting that empirical likelihood techniques could be well-suited to this setting. And (ii) it is no longer a mathematical worry: there exist a joint distribution on m(x,θ), induced by a (or many) joint distribution on (x,θ). So the question of finding whether or not it induces a single proper prior on θ becomes relevant. But, if I want to use ABC, being given the distribution of m(x,θ) seems to mean I can only generate new values of this transform while missing a natural distance between observations and pseudo-observations. Still, I entertain lingering doubts that this is the meaning of the study. Where does the joint distribution come from..?!

“Typically C is coarse in the sense that it does not contain all the Borel sets (…)  The probability space cannot be used for Bayesian inference”

My understanding of that part is that defining a joint on m(x,θ) is not always enough to deduce a (unique) posterior on θ, which is fine and correct, but rather anticlimactic. This sounds to be what Gallant calls a “partial specification of the prior” (p.9).

Overall, after this linear read, I remain very much puzzled by the statistical (or Bayesian) implications of the paper . The fact that the moment conditions are central to the approach would once again induce me to check the properties of an alternative approach like empirical likelihood.

## an ABC experiment

Posted in Books, pictures, R, Statistics, University life with tags , , , , , , , , on November 24, 2014 by xi'an

In a cross-validated forum exchange, I used the code below to illustrate the working of an ABC algorithm:

#normal data with 100 observations
n=100
x=rnorm(n)
#observed summaries

#normal x gamma prior
priori=function(N){
return(cbind(rnorm(N,sd=10),
1/sqrt(rgamma(N,shape=2,scale=5))))
}

ABC=function(N,alpha=.05){

prior=priori(N) #reference table

#pseudo-data
summ=matrix(0,N,2)
for (i in 1:N){
xi=rnorm(n)*prior[i,2]+prior[i,1]
}

#normalisation factor for the distance
#distance
#selection
posterior=prior[dist<quantile(dist,alpha),]}


Hence I used the median and the mad as my summary statistics. And the outcome is rather surprising, for two reasons: the first one is that the posterior on the mean μ is much wider than when using the mean and the variance as summary statistics. This is not completely surprising in that the latter are sufficient, while the former are not. Still, the (-10,10) range on the mean is way larger… The second reason for surprise is that the true posterior distribution cannot be derived since the joint density of med and mad is unavailable.

After thinking about this for a while, I went back to my workbench to check the difference with using mean and variance. To my greater surprise, I found hardly any difference! Using the almost exact ABC with 10⁶ simulations and a 5% subsampling rate returns exactly the same outcome. (The first row above is for the sufficient statistics (mean,standard deviation) while the second row is for the (median,mad) pair.) Playing with the distance does not help. The genuine posterior output is quite different, as exposed on the last row of the above, using a basic Gibbs sampler since the posterior is not truly conjugate.

## Challis Lectures

Posted in Books, pictures, Statistics, Travel, University life, Wines with tags , , , , , , , on November 23, 2014 by xi'an

I had a great time during this short visit in the Department of Statistics, University of Florida, Gainesville. First, it was a major honour to be the 2014 recipient of the George H. Challis Award and I considerably enjoyed delivering my lectures on mixtures and on ABC with random forests, And chatting with members of the audience about the contents afterwards. Here is the physical award I brought back to my office:

More as a piece of trivia, here is the amount of information about the George H. Challis Award I found on the UF website:

This fund was established in 2000 by Jack M. and Linda Challis Gill and the Gill Foundation of Texas, in memory of Linda’s father, to support faculty and student conference travel awards and the George Challis Biostatistics Lecture Series. George H. Challis was born on December 8, 1911 and was raised in Italy and Indiana. He was the first cousin of Indiana composer Cole Porter. George earned a degree in 1933 from the School of Business at Indiana University in Bloomington. George passed away on May 6, 2000. His wife, Madeline, passed away on December 14, 2009.

Cole Porter, indeed!

On top of this lecturing activity, I had a full academic agenda, discussing with most faculty members and PhD students of the Department, on our respective research themes over the two days I was there and it felt like there was not enough time! And then, during the few remaining hours where I did not try to stay on French time (!), I had a great time with my friends Jim and Maria in Gainesville, tasting a fantastic local IPA beer from Cigar City Brewery and several great (non-local) red wines… Adding to that a pile of new books, a smooth trip both ways, and a chance encounter with Alicia in Atlanta airport, it was a brilliant extended weekend!