In connection with our handbook on mixtures being published, here are three chapters I contributed to from the Handbook of ABC, edited by Scott Sisson, Yanan Fan, and Mark Beaumont:
6. Likelihood-free Model Choice, by J.-M. Marin, P. Pudlo, A. Estoup and C.P. Robert
12. Approximating the Likelihood in ABC, by C. C. Drovandi, C. Grazian, K. Mengersen and C.P. Robert
A newly arXived paper from a group of researchers at NUS I wish we had discussed when I was there last month. As we wrote this empirical ABCe paper in PNAS with Kerrie Mengersen and Pierre Pudlo in 2012. Plus the SAME paper with Arnaud Doucet and Simon Godsill ten years earlier, which the authors prefer to call data cloning in continuation of the more recent Lele et al. (2007). They could actually have used my original denomination of prior feedback (1992? I remember presenting the idea at Camp Casella in Cornell that summer) as well! Actually, I am not certain invoking prior feedback is quite necessary since this is a form of simulated method of moments as well.
Now, did we really assume that some moments of the distribution were analytically available, although the likelihood was not?! Even before going through the paper, it dawned on me that these theoretical moments could have been simulated instead, since the model is a generative one: for a given parameter value, a direct Monte Carlo approximation to the exact moment can be produced and can serve as a constraint for the empirical likelihood definition. I am surprised and aggrieved that we would not think of this empirical likelihood version of a method of moments. Which is central to the current paper. In the sense that, were the parameter exact, the differences between the moments based on the actual data x⁰ and the moments based on m replicas of the simulated data x¹,x²,… have mean zero, meaning the moment constraint is immediately available. Meaning an empirical likelihood is easily constructed, replacing the actual likelihood in an MCMC scheme, albeit at a rather high computing cost. Congratulations to the authors for uncovering this possibility that we missed!
“The summary statistics in this example were judiciously chosen.”
One point in the paper on which I disagree with the authors is the argument that MCMC sampling based on an empirical likelihood can be seen as an implementation of the pseudo-marginal Metropolis-Hastings method. The major difference in my opinion is that there is no unbiasedness here (and no generic result that indicates convergence to the exact posterior as the number of simulations grows to infinity). The other point unclear to me is about the selection of summaries [or moments] for implementing the method, which seems to be based on their performances in the subsequent estimation, performances that are hard to assess properly in intractable likelihood cases. In the last example of stereological extremes (not covered in our paper), for instance, the output is compared with the parallel synthetic likelihood result.
While working on the Series B’log the other day I noticed this paper by Chauduri et al. on Hamiltonian Monte Carlo and empirical likelihood: how exciting!!! Here is the abstract of the paper:
We consider Bayesian empirical likelihood estimation and develop an efficient Hamiltonian Monte Car lo method for sampling from the posterior distribution of the parameters of interest.The method proposed uses hitherto unknown properties of the gradient of the underlying log-empirical-likelihood function. We use results from convex analysis to show that these properties hold under minimal assumptions on the parameter space, prior density and the functions used in the estimating equations determining the empirical likelihood. Our method employs a finite number of estimating equations and observations but produces valid semi-parametric inference for a large class of statistical models including mixed effects models, generalized linear models and hierarchical Bayes models. We overcome major challenges posed by complex, non-convex boundaries of the support routinely observed for empirical likelihood which prevent efficient implementation of traditional Markov chain Monte Car lo methods like random-walk Metropolis–Hastings sampling etc. with or without parallel tempering. A simulation study confirms that our method converges quickly and draws samples from the posterior support efficiently. We further illustrate its utility through an analysis of a discrete data set in small area estimation.
[The comment is reposted from Series B’log, where I wrote it first.]
It is of particular interest for me [disclaimer: I was not involved in the review of this paper!] as we worked on ABC thru empirical likelihood, which is about the reverse of the current paper in terms of motivation: when faced with a complex model, we substitute an empirical likelihood version for the real thing, run simulations from the prior distribution and use the empirical likelihood as a proxy. With possible intricacies when the data is not iid (an issue we also met with Wasserstein distances.) In this paper the authors instead consider working on an empirical likelihood as their starting point and derive an HMC algorithm to do so. The idea is striking in that, by nature, an empirical likelihood is not a very smooth object and hence does not seem open to producing gradients and Hessians. As illustrated by Figure 1 in the paper . Which is so spiky at places that one may wonder at the representativity of such graphs.
I have always had a persistent worry about the ultimate validity of treating the empirical likelihood as a genuine likelihood, from the fact that it is the result of an optimisation problem to the issue that the approximate empirical distribution has a finite (data-dependent) support, hence is completely orthogonal to the true distribution. And to the one that the likelihood function is zero outside the convex hull of the defining equations…(For one thing, this empirical likelihood is always bounded by one but this may be irrelevant after all!)
The computational difficulty in handling the empirical likelihood starts with its support. Eliminating values of the parameter for which this empirical likelihood is zero amounts to checking whether zero belongs to the above convex hull. A hard (NP hard?) problem. (Although I do not understand why the authors dismiss the token observations of Owen and others. The argument that Bayesian analysis does more than maximising a likelihood seems to confuse the empirical likelihood as a product of a maximisation step with the empirical likelihood as a function of the parameter that can be used as any other function.)
In the simple regression example (pp.297-299), I find the choice of the moment constraints puzzling, in that they address the mean of the white noise (zero) and the covariance with the regressors (zero too). Puzzling because my definition of the regression model is conditional on the regressors and hence does not imply anything on their distribution. In a sense this is another model. But I also note that the approach focus on the distribution of the reconstituted white noises, as we did in the PNAS paper. (The three examples processed in the paper are all simple and could be processed by regular MCMC, thus making the preliminary step of calling for an empirical likelihood somewhat artificial unless I missed the motivation. The paper also does not seem to discuss the impact of the choice of the moment constraints or the computing constraints involved by a function that is itself the result of a maximisation problem.)
A significant part of the paper is dedicated to the optimisation problem and the exclusion of the points on the boundary. Which sounds like a non-problem in continuous settings. However, this appears to be of importance for running an HMC as it cannot evade the support (without token observations). On principle, HMC should not leave this support since the gradient diverges at the boundary, but in practice the leapfrog approximation may lead the path outside. I would have (naïvely?) suggested to reject moves when this happens and start again but the authors consider that proper choices of the calibration factors of HMC can avoid this problem. Which seems to induce a practical issue by turning the algorithm into an adaptive version.
As a last point, I would have enjoyed seeing a comparison of the performances against our (A)BCel version, which would have been straightforward to implement in the simple examples handled by the paper. (This could be a neat undergraduate project for next year!)
This recently arXived paper by Weixuan Zhu , Juan Miguel Marín, and Fabrizio Leisen proposes an alternative to our empirical likelihood ABC paper of 2013, or BCel. Besides the mostly personal appeal for me to report on a Juan Miguel Marín working [in Madrid] on ABC topics, along my friend Jean-Michel Marin!, this paper is another entry on ABC that connects with yet another statistical perspective, namely bootstrap. The proposal, called BCbl, is based on a reference paper by Davison, Hinkley and Worton (1992) which defines a bootstrap likelihood, a notion that relies on a double-bootstrap step to produce a non-parametric estimate of the distribution of a given estimator of the parameter θ. This estimate includes a smooth curve-fitting algorithm step, for which little description is available from the current paper. The bootstrap non-parametric substitute then plays the role of the actual likelihood, with no correction for the substitution just as in our BCel. Both approaches are convergent, with Monte Carlo simulations exhibiting similar or even identical convergence speeds although [unsurprisingly!] no deep theory is available on the comparative advantage.
An important issue from my perspective is that, while the empirical likelihood approach relies on a choice of identifying constraints that strongly impact the numerical value of the likelihood approximation, the bootstrap version starts directly from a subjectively chosen estimator of θ, which may also impact the numerical value of the likelihood approximation. In some ABC settings, finding a primary estimator of θ may be a real issue or a computational burden. Except when using a preliminary ABC step as in semi-automatic ABC. This would be an interesting crash-test for the BCbl proposal! (This would not necessarily increase the computational cost by a large amount.) In addition, I am not sure the method easily extends to larger collections of summary statistics as those used in ABC, in particular because it necessarily relies on non-parametric estimates, only operating in small enough dimensions where smooth curve-fitting algorithms can be used. Critically, the paper only processes examples with a few parameters.
The comparisons between BCel and BCbl that are produced in the paper show some gain towards BCbl. Obviously, it depends on the respective calibrations of the non-parametric methods and of regular ABC, as well as on the available computing time. I find the population genetic example somewhat puzzling: The paper refers to our composite likelihood to set the moment equations. Since this is a pseudo-likelihood, I wonder how the authors do select their parameter estimates in the double-bootstrap experiment. And for the Ising model, it is not straightforward to conceive of a bootstrap algorithm on an Ising model: (a) how does one subsample pixels and (b) what are the validity guarantees for the estimation procedure.
This week in Warwick was one of the busiest ones ever as I had to juggle between two workshops, including one in Oxford, a departmental meeting, two paper revisions, two pre-vivas, and a seminar in Leeds. Not to mention a broken toe (!), a flat tire (!!), and a diner at the X. Hardly anytime for writing blog entries..! Fortunately, I managed to squeeze time for working with Kerrie Mengersen who was visiting Warwick this fortnight. Finding new directions for the (A)BCel approach we developed a few years ago with Pierre Pudlo. The workshop in Oxford was quite informal with talks from PhD students [I fear I cannot discuss here as the papers are not online yet]. And one talk by François Caron about estimating sparse networks with not exactly exchangeable priors and completely random measures. And one talk by Kerrie Mengersen on a new and in-progress approach to handling Big Data that I found quite convincing (if again cannot discuss here). The probabilistic numerics workshop was discussed in yesterday’s post and I managed to discuss it a wee bit further with the organisers at The X restaurant in Kenilworth. (As a superfluous aside, and after a second sampling this year, I concluded that the Michelin star somewhat undeserved in that the dishes at The X are not particularly imaginative or tasty, the excellent sourdough bread being the best part of the meal!) I was expecting the train ride to Leeds to be highly bucolic as it went through the sunny countryside of South Yorkshire, with newly born lambs running in the bright green fields surrounded by old stone walls…, but instead went through endless villages with their rows of brick houses. Not that I have anything against brick houses, mind! Only, I had not realised how dense this part of England was, this presumably getting back all the way to the Industrial Revolution with the Manchester-Leeds-Birmingham triangle.
My seminar in Leeds was as exciting as in Amsterdam last week and with a large audience, so I got many and only interesting questions, from the issue of turning the output (i.e., the posterior on α) into a decision rule, to making decision in the event of a non-conclusive posterior, to links with earlier frequentist resolutions, to whether or not we were able to solve the Lindley-Jeffreys paradox (we are not!, which makes a lot of sense), to the possibility of running a subjective or a sequential version. After the seminar I enjoyed a perfect Indian dinner at Aagrah, apparently a Yorkshire institution, with the right balance between too hot and too mild, i.e., enough spices to break a good sweat but not too many to loose any sense of taste!
Xi'an's Og 