Archive for Monte Carlo Statistical Methods

ergodicity of approximate MCMC chains with applications to large datasets

Posted in pictures, Statistics, Travel, University life with tags , , , , , , , , , , on August 31, 2015 by xi'an

bhamAnother arXived paper I read on my way to Warwick! And yet another paper written by my friend Natesh Pillai (and his co-author Aaron Smith, from Ottawa). The goal of the paper is to study the ergodicity and the degree of approximation of the true posterior distribution of approximate MCMC algorithms that recently flourished as an answer to “Big Data” issues… [Comments below are about the second version of this paper.] One of the most curious results in the paper is the fact that the approximation may prove better than the original kernel, in terms of computing costs! If asymptotically in the computing cost. There also are acknowledged connections with the approximative MCMC kernel of Pierre Alquier, Neal Friel, Richard Everitt and A Boland, briefly mentioned in an earlier post.

The paper starts with a fairly theoretical part, to follow with an application to austerity sampling [and, in the earlier version of the paper, to the Hoeffding bounds of Bardenet et al., both discussed earlier on the ‘Og, to exponential random graphs (the paper being rather terse on the description of the subsampling mechanism), to stochastic gradient Langevin dynamics (by Max Welling and Yee-Whye Teh), and to ABC-MCMC]. The assumptions are about the transition kernels of a reference Markov kernel and of one associated with the approximation, imposing some bounds on the Wasserstein distance between those kernels, K and K’. Results being generic, there is no constraint as to how K is chosen or on how K’ is derived from K. Except in Lemma 3.6 and in the application section, where the same proposal kernel L is used for both Metropolis-Hastings algorithms K and K’. While I understand this makes for an easier coupling of the kernels, this also sounds like a restriction to me in that modifying the target begs for a similar modification in the proposal, if only because the tails they are a-changin’

In the case of subsampling the likelihood to gain computation time (as discussed by Korattikara et al. and by Bardenet et al.), the austerity algorithm as described in Algorithm 2 is surprising as the average of the sampled data log-densities and the log-transform of the remainder of the Metropolis-Hastings probability, which seem unrelated, are compared until they are close enough.  I also find hard to derive from the different approximation theorems bounding exceedance probabilities a rule to decide on the subsampling rate as a function of the overall sample size and of the computing cost. (As a side if general remark, I remain somewhat reserved about the subsampling idea, given that it requires the entire dataset to be available at every iteration. This makes parallel implementations rather difficult to contemplate.)

JSM 2015 [day #4]

Posted in pictures, Running, Statistics, Travel, University life with tags , , , , , , , , , , , , , , , , , , on August 13, 2015 by xi'an

My first session today was Markov Chain Monte Carlo for Contemporary Statistical Applications with a heap of interesting directions in MCMC research! Now, without any possible bias (!), I would definitely nominate Murray Pollock (incidentally from Warwick) as the winner for best slides, funniest presentation, and most enjoyable accent! More seriously, the scalable Langevin algorithm he developed with Paul Fearnhead, Adam Johansen, and Gareth Roberts, is quite impressive in avoiding computing costly likelihoods. With of course caveats on which targets it applies to. Murali Haran showed a new proposal to handle high dimension random effect models by a projection trick that reduces the dimension. Natesh Pillai introduced us (or at least me!) to a spectral clustering that allowed for an automated partition of the target space, itself the starting point to his parallel MCMC algorithm. Quite exciting, even though I do not perceive partitions as an ideal solution to this problem. The final talk in the session was Galin Jones’ presentation of consistency results and conditions for multivariate quantities which is a surprisingly unexplored domain. MCMC is still alive and running!

The second MCMC session of the morning, Monte Carlo Methods Facing New Challenges in Statistics and Science, was equally diverse, with Lynn Kuo’s talk on the HAWK approach, where we discovered that harmonic mean estimators are still in use, e.g., in MrBayes software employed in phylogenetic inference. The proposal to replace this awful estimator that should never be seen again (!) was rather closely related to an earlier solution of us for marginal likelihood approximation, based there on a partition of the whole space rather than an HPD region in our case… Then, Michael Betancourt brilliantly acted as a proxy for Andrew to present the STAN language, with a flashy trailer he most recently designed. Featuring Andrew as the sole actor. And with great arguments for using it, including the potential to run expectation propagation (as a way of life). In fine, Faming Liang proposed a bootstrap subsampling version of the Metropolis-Hastings algorithm, where the likelihood acknowledging the resulting bias in the limiting distribution.

My first afternoon session was another entry on Statistical Phylogenetics, somewhat continued from yesterday’s session. Making me realised I had not seen a single talk on ABC for the entire meeting! The issues discussed in the session were linked with aligning sequences and comparing  many trees. Again in settings where likelihoods can be computed more or less explicitly. Without any expertise in the matter, I wondered at a construction that would turn all trees, like  into realizations of a continuous model. For instance by growing one branch at a time while removing the MRCA root… And maybe using a particle like method to grow trees. As an aside, Vladimir Minin told me yesterday night about genetic mutations that could switch on and off phenotypes repeatedly across generations… For instance  the ability to glow in the dark for species of deep sea fish.

When stating that I did not see a single talk about ABC, I omitted Steve Fienberg’s Fisher Lecture R.A. Fisher and the Statistical ABCs, keeping the morceau de choix for the end! Even though of course Steve did not mention the algorithm! A was for asymptotics, or ancilarity, B for Bayesian (or biducial??), C for causation (or cuffiency???)… Among other germs, I appreciated that Steve mentioned my great-grand father Darmois in connection with exponential families! And the connection with Jon Wellner’s LeCam Lecture from a few days ago. And reminding us that Savage was a Fisher lecturer himself. And that Fisher introduced fiducial distributions quite early. And for defending the Bayesian perspective. Steve also set some challenges like asymptotics for networks, Bayesian model assessment (I liked the notion of stepping out of the model), and randomization when experimenting with networks. And for big data issues. And for personalized medicine, building on his cancer treatment. No trace of the ABC algorithm, obviously, but a wonderful Fisher’s lecture, also most obviously!! Bravo, Steve, keep thriving!!!

JSM 2015 [day #1]

Posted in Books, R, Statistics, Travel, University life with tags , , , , , , , , , , , , , , , on August 10, 2015 by xi'an

ferryThis afternoon, at JSM 2015, in Seattle, we had the Bayesian Computation I and II sessions that Omiros Papaspiliopoulos and myself put together (sponsored by IMS and ISBA). Despite this being Sunday and hence having some of the participants still arriving, the sessions went on well in terms of audience. Thanks to Mark Girolami’s strict presidency, we were so much on time in Bayesian Computation I that we had 20mn left for a floor discussion that turned into a speakers’ discussion! All talks were of obvious interest for MCMCists, but Ryan Adams’ presentation on firefly Monte Carlo got me thinking for most of the afternoon on different ways of exploiting the existence of a bound on the terms composing the target. With little to show by the end of the afternoon! On the mundane side, I was sorry to miss Pierre Jacob, who was still in France due to difficulties in obtaining a working visa for Harvard (!), and surprised to see Dawn Woodard wearing a Uber tee-shirt, until she told us she was now working at Uber! Which a posteriori makes sense, given her work on traffic predictions!

delayed in Seattle

Posted in Books, R, Statistics, Travel, University life with tags , , , , , , , , , , on August 9, 2015 by xi'an

Here are the slides of my talk on delayed acceptance I present this afternoon at JSM 2015, in Seattle, in the Bayesian Computation I (2pm, room CC-4C1) and II (4pm, room CC-3A) sessions Omiros Papaspiliopoulos and myself put together (sponsored by IMS and ISBA):

on estimating constants…

Posted in pictures, Statistics, Travel, University life with tags , , , , , , , , , on July 21, 2015 by xi'an

While I discussed on the ‘Og in the past the difference I saw between estimating an unknown parameter from a distribution and evaluating a normalising constant, evaluating such constants and hence handling [properly] doubly intractable models is obviously of the utmost importance! For this reason, Nial Friel, Helen Ogden and myself have put together a CRiSM workshop on the topic (with the tongue-in-cheek title of Estimating constants!), to be held at the University of Warwick next April 20-22.

The CRiSM workshop will focus on computational methods for approximating challenging normalising constants found in Monte Carlo, likelihood and Bayesian models. Such methods may be used in a wide range of problems: to compute intractable likelihoods, to find the evidence in Bayesian model selection, and to compute the partition function in Physics. The meeting will bring together different communities working on these related problems, some of which have developed original if little advertised solutions. It will also highlight the novel challenges associated with large data and highly complex models. Besides a dozen invited talks, the schedule will highlight two afternoon poster sessions with speed (2-5mn) oral presentations called ‘Elevator’ talks.

While 2016 is going to be quite busy with all kinds of meetings (MCMSkv, ISBA 2016, the CIRM Statistics month, AISTATS 2016, …), this should be an exciting two-day workshop, given the on-going activity in this area, and I thus suggest interested readers to mark the dates in their diary. I will obviously keep you posted about registration and accommodation when those entries are available.

MCMskv, Lenzerheide, 4-7 Jan., 2016 [news #1]

Posted in Kids, Mountains, pictures, R, Statistics, Travel, University life with tags , , , , , , , , , , , , , , , , , , on July 20, 2015 by xi'an

moonriseThe BayesComp MCMski V [or MCMskv for short] has now its official website, once again maintained by Merrill Lietchy from Drexel University, Philadelphia, and registration is even open! The call for contributed sessions is now over, while the call for posters remains open until the very end. The novelty from the previous post is that there will be a “Breaking news” [in-between the Late news sessions at JSM and the crash poster talks at machine-learning conferences] session to highlight major advances among poster submissions. And that there will be an opening talk by Steve [the Bayesian] Scott on the 4th, about the frightening prospect of MCMC death!, followed by a round-table and a welcome reception, sponsored by the Swiss Supercomputing Centre. Hence the change in dates. Which still allows for arrivals in Zürich on the January 4th [be with you].

Leave the Pima Indians alone!

Posted in Books, R, Statistics, University life with tags , , , , , , , , , , , , , , , , , on July 15, 2015 by xi'an

“…our findings shall lead to us be critical of certain current practices. Specifically, most papers seem content with comparing some new algorithm with Gibbs sampling, on a few small datasets, such as the well-known Pima Indians diabetes dataset (8 covariates). But we shall see that, for such datasets, approaches that are even more basic than Gibbs sampling are actually hard to beat. In other words, datasets considered in the literature may be too toy-like to be used as a relevant benchmark. On the other hand, if ones considers larger datasets (with say 100 covariates), then not so many approaches seem to remain competitive” (p.1)

Nicolas Chopin and James Ridgway (CREST, Paris) completed and arXived a paper they had “threatened” to publish for a while now, namely why using the Pima Indian R logistic or probit regression benchmark for checking a computational algorithm is not such a great idea! Given that I am definitely guilty of such a sin (in papers not reported in the survey), I was quite eager to read the reasons why! Beyond the debate on the worth of such a benchmark, the paper considers a wider perspective as to how Bayesian computation algorithms should be compared, including the murky waters of CPU time versus designer or programmer time. Which plays against most MCMC sampler.

As a first entry, Nicolas and James point out that the MAP can be derived by standard a Newton-Raphson algorithm when the prior is Gaussian, and even when the prior is Cauchy as it seems most datasets allow for Newton-Raphson convergence. As well as the Hessian. We actually took advantage of this property in our comparison of evidence approximations published in the Festschrift for Jim Berger. Where we also noticed the awesome performances of an importance sampler based on the Gaussian or Laplace approximation. The authors call this proposal their gold standard. Because they also find it hard to beat. They also pursue this approximation to its logical (?) end by proposing an evidence approximation based on the above and Chib’s formula. Two close approximations are provided by INLA for posterior marginals and by a Laplace-EM for a Cauchy prior. Unsurprisingly, the expectation-propagation (EP) approach is also implemented. What EP lacks in theoretical backup, it seems to recover in sheer precision (in the examples analysed in the paper). And unsurprisingly as well the paper includes a randomised quasi-Monte Carlo version of the Gaussian importance sampler. (The authors report that “the improvement brought by RQMC varies strongly across datasets” without elaborating for the reasons behind this variability. They also do not report the CPU time of the IS-QMC, maybe identical to the one for the regular importance sampling.) Maybe more surprising is the absence of a nested sampling version.

pimcisIn the Markov chain Monte Carlo solutions, Nicolas and James compare Gibbs, Metropolis-Hastings, Hamiltonian Monte Carlo, and NUTS. Plus a tempering SMC, All of which are outperformed by importance sampling for small enough datasets. But get back to competing grounds for large enough ones, since importance sampling then fails.

“…let’s all refrain from now on from using datasets and models that are too simple to serve as a reasonable benchmark.” (p.25)

This is a very nice survey on the theme of binary data (more than on the comparison of algorithms in that the authors do not really take into account design and complexity, but resort to MSEs versus CPus). I however do not agree with their overall message to leave the Pima Indians alone. Or at least not for the reason provided therein, namely that faster and more accurate approximations methods are available and cannot be beaten. Benchmarks always have the limitation of “what you get is what you see”, i.e., the output associated with a single dataset that only has that many idiosyncrasies. Plus, the closeness to a perfect normal posterior makes the logistic posterior too regular to pause a real challenge (even though MCMC algorithms are as usual slower than iid sampling). But having faster and more precise resolutions should on the opposite be  cause for cheers, as this provides a reference value, a golden standard, to check against. In a sense, for every Monte Carlo method, there is a much better answer, namely the exact value of the integral or of the optimum! And one is hardly aiming at a more precise inference for the benchmark itself: those Pima Indians [whose actual name is Akimel O’odham] with diabetes involved in the original study are definitely beyond help from statisticians and the model is unlikely to carry out to current populations. When the goal is to compare methods, as in our 2009 paper for Jim Berger’s 60th birthday, what matters is relative speed and relative ease of implementation (besides the obvious convergence to the proper target). In that sense bigger and larger is not always relevant. Unless one tackles really big or really large datasets, for which there is neither benchmark method nor reference value.


Get every new post delivered to your Inbox.

Join 905 other followers