flow contrastive estimation

On the flight back from Montpellier, last week, I read a 2019 paper by Gao et al. revisiting the MLE estimation of a parametric family parameter when the normalising constant Z=Z(θ) is unknown. Via noise-contrastive estimation à la Guttman & Hyvärinnen (or à la Charlie Geyer). Treating the normalising constant Z as an extra parameter (as in Kong et al.) and the classification probability as an objective function and calling it a likelihood, which it is not in my opinion as (i) the allocation to the groups is not random and (ii) the original density of the actual observations does not appear in the so-called likelihood.

“When q appears on the right of KL-divergence [against p],  it is forced to cover most of the modes of p, When q appears on the left of KL-divergence, it tends to chase the major modes of p while ignoring the minor modes.”

The flow in the title indicates that the contrastive distribution q is estimated by a flow-based estimator, namely the transform of a basic noise distribution via easily invertible and differentiable transforms, for instance with lower triangular Jacobians. This flow is also estimated directly from the data but the authors complain this estimation is not good enough for noise contrastive estimation and suggest instead resorting to a GAN version where the classification log-probability is maximised in the model parameters and minimsed in the flow parameters. Except that I feel it misses the true likelihood part. In other words, why on Hyperion would estimating all θ, Z=Z(θ), and α at once improve the estimation of Z?

The other aspect that puzzles me is that (12) uses integrated classification probabilities (with the unknown Z as extra parameter), rather than conditioning on the data, Bayes-like. (The difference between (12) and GAN is that here the discriminator function is constrained.) Esp. when the first expectation is replaced with its empirical version.

marginal likelihood as exhaustive X validation

In the June issue of Biometrika (for which I am deputy editor) Edwin Fong and Chris Holmes have a short paper (that I did not process!) on the validation of the marginal likelihood as the unique coherent updating rule. Marginal in the general sense of Bissiri et al. (2016). Coherent in the sense of being invariant to the order of input of exchangeable data, if in a somewhat self-defining version (Definition 1). As a consequence, marginal likelihood arises as the unique prequential scoring rule under coherent belief updating in the Bayesian framework. (It is unique given the prior or its generalisation, obviously.)

“…we see that 10% of terms contributing to the marginal likelihood come from out-of-sample predictions, using on average less than 5% of the available training data.”

The paper also contains the interesting remark that the log marginal likelihood is the average leave-p-out X-validation score, across all values of p. Which shows that, provided the marginal can be approximated, the X validation assessment is feasible. Which leads to a highly relevant (imho) spotlight on how this expresses the (deadly) impact of the prior selection on the numerical value of the marginal likelihood. Leaving outsome of the least informative terms in the X-validation leads to exactly the log geometric intrinsic Bayes factor of Berger & Pericchi (1996). Most interesting connection with the Bayes factor community but one that depends on the choice of the dismissed fraction of p‘s.

a stretched view on Keynes’ Treatise

I came across a rather bemusing interpretation of Keynes’ Treatise on Probability, as a tribune in Le Monde of 6 September, as being a statement against the mathematical modelling of economy. Written by Annie Cot, professor of economics at Paris Sorbonne University. While the philosophical thread of the book is inclined towards a subjective perception of probability, albeit rejecting the Bayesian approach, and while the view on statistics is equally pessimistic, falling into the infinite regress of conditioning on the observation itself, outside a Bayesian framework, as I discussed in my 2011 paper, the book makes no mention whatsoever of economics or economic models. As far as I remember the book from reading it ten years ago. To conclude, as the author of this tribune, that Keynes rejected the viability of prevision based on economic models via this book sounds therefore stretching the facts to a fair extent.

sequential neural likelihood estimation as ABC substitute

A JMLR paper by Papamakarios, Sterratt, and Murray (Edinburgh), first presented at the AISTATS 2019 meeting, on a new form of likelihood-free inference, away from non-zero tolerance and from the distance-based versions of ABC, following earlier papers by Iain Murray and co-authors in the same spirit. Which I got pointed to during the ABC workshop in Vancouver. At the time I had no idea as to autoregressive flows meant. We were supposed to hold a reading group in Paris-Dauphine on this paper last week, unfortunately cancelled as a coronaviral precaution… Here are some notes I had prepared for the meeting that did not take place.

“A simulator model is a computer program, which takes a vector of parameters θ, makes internal calls to a random number generator, and outputs a data vector x.”

Just the usual generative model then.

“A conditional neural density estimator is a parametric model q(.|φ) (such as a neural network) controlled by a set of parameters φ, which takes a pair of datapoints (u,v) and outputs a conditional probability density q(u|v,φ).”

Less usual, in that the outcome is guaranteed to be a probability density.

“For its neural density estimator, SNPE uses a Mixture Density Network, which is a feed-forward neural network that takes x as input and outputs the parameters of a Gaussian mixture over θ.”

In which theoretical sense would it improve upon classical or Bayesian density estimators? Where are the error evaluation, the optimal rates, the sensitivity to the dimension of the data? of the parameter?

“Our new method, Sequential Neural Likelihood (SNL), avoids the bias introduced by the proposal, by opting to learn a model of the likelihood instead of the posterior.”

I do not get the argument in that the final outcome (of using the approximation within an MCMC scheme) remains biased since the likelihood is not the exact likelihood. Where is the error evaluation? Note that in the associated Algorithm 1, the learning set is enlarged on each round, as in AMIS, rather than set back to the empty set ∅ on each round.

“…given enough simulations, a sufficiently flexible conditional neural density estimator will eventually approximate the likelihood in the support of the proposal, regardless of the shape of the proposal. In other words, as long as we do not exclude parts of the parameter space, the way we propose parameters does not bias learning the likelihood asymptotically. Unlike when learning the posterior, no adjustment is necessary to account for our proposing strategy.”

This is a rather vague statement, with the only support being that the Monte Carlo approximation to the Kullback-Leibler divergence does converge to its actual value, i.e. a direct application of the Law of Large Numbers! But an interesting point I informally made a (long) while ago that all that matters is the estimate of the density at x⁰. Or at the value of the statistic at x⁰. The masked auto-encoder density estimator is based on a sequence of bijections with a lower-triangular Jacobian matrix, meaning the conditional density estimate is available in closed form. Which makes it sounds like a form of neurotic variational Bayes solution.

The paper also links with ABC (too costly?), other parametric approximations to the posterior (like Gaussian copulas and variational likelihood-free inference), synthetic likelihood, Gaussian processes, noise contrastive estimation… With experiments involving some of the above. But the experiments involve rather smooth models with relatively few parameters.

“A general question is whether it is preferable to learn the posterior or the likelihood (…) Learning the likelihood can often be easier than learning the posterior, and it does not depend on the choice of proposal, which makes learning easier and more robust (…) On the other hand, methods such as SNPE return a parametric model of the posterior directly, whereas a further inference step (e.g. variational inference or MCMC) is needed on top of SNL to obtain a posterior estimate”

A fair point in the conclusion. Which also mentions the curse of dimensionality (both for parameters and observations) and the possibility to work directly with summaries.

Getting back to the earlier and connected Masked autoregressive flow for density estimation paper, by Papamakarios, Pavlakou and Murray:

“Viewing an autoregressive model as a normalizing flow opens the possibility of increasing its flexibility by stacking multiple models of the same type, by having each model provide the source of randomness for the next model in the stack. The resulting stack of models is a normalizing flow that is more flexible than the original model, and that remains tractable.”

Which makes it sound like a sort of a neural network in the density space. Optimised by Kullback-Leibler minimisation to get asymptotically close to the likelihood. But a form of Bayesian indirect inference in the end, namely an MLE on a pseudo-model, using the estimated model as a proxy in Bayesian inference…

What the …?!