arXiv subjects

Nicolas Chopin

Publications and source records attributed to Nicolas Chopin.

At least 37 records · Page 2Linked to original sources

arXiv2023

Scalable Bayesian bi-level variable selection in generalized linear models

Motivated by a real-world application in cardiology, we develop an algorithm to perform Bayesian bi-level variable selection in a generalized linear model, for datasets that may be large both in terms of the number of individuals and the number of predictors. Our algorithm relies on the waste-free SMC Sequential Monte Carlo methodology of Dau and Chopin (2022), a new proposal mechanism to deal with the constraints specific to bi-level selection (which forbid to select an individual predictor if its group is not selected), and the ALA (approximate Laplace approximation) approach of Rossell et al. (2021). We show in our numerical study that the algorithm may offer reliable performance on large datasets within a few minutes, on both simulated data and real data related to the aforementioned cardiology application.

stat.ME↗

arXiv2023

On backward smoothing algorithms

In the context of state-space models, skeleton-based smoothing algorithms rely on a backward sampling step which by default has a $\mathcal O(N^2)$ complexity (where $N$ is the number of particles). Existing improvements in the literature are unsatisfactory: a popular rejection sampling -- based approach, as we shall show, might lead to badly behaved execution time; another rejection sampler with stopping lacks complexity analysis; yet another MCMC-inspired algorithm comes with no stability guarantee. We provide several results that close these gaps. In particular, we prove a novel non-asymptotic stability theorem, thus enabling smoothing with truly linear complexity and adequate theoretical justification. We propose a general framework which unites most skeleton-based smoothing algorithms in the literature and allows to simultaneously prove their convergence and stability, both in online and offline contexts. Furthermore, we derive, as a special case of that framework, a new coupling-based smoothing algorithm applicable to models with intractable transition densities. We elaborate practical recommendations and confirm those with numerical experiments.

stat.CO↗

arXiv2022

On resampling schemes for particle filters with weakly informative observations

We consider particle filters with weakly informative observations (or `potentials') relative to the latent state dynamics. The particular focus of this work is on particle filters to approximate time-discretisations of continuous-time Feynman--Kac path integral models -- a scenario that naturally arises when addressing filtering and smoothing problems in continuous time -- but our findings are indicative about weakly informative settings beyond this context too. We study the performance of different resampling schemes, such as systematic resampling, SSP (Srinivasan sampling process) and stratified resampling, as the time-discretisation becomes finer and also identify their continuous-time limit, which is expressed as a suitably defined `infinitesimal generator.' By contrasting these generators, we find that (certain modifications of) systematic and SSP resampling `dominate' stratified and independent `killing' resampling in terms of their limiting overall resampling rate. The reduced intensity of resampling manifests itself in lower variance in our numerical experiment. This efficiency result, through an ordering of the resampling rate, is new to the literature. The second major contribution of this work concerns the analysis of the limiting behaviour of the entire population of particles of the particle filter as the time discretisation becomes finer. We provide the first proof, under general conditions, that the particle approximation of the discretised continuous-time Feynman--Kac path integral models converges to a (uniformly weighted) continuous-time particle system.

stat.CO↗

arXiv2022

De-biasing particle filtering for a continuous time hidden Markov model with a Cox process observation model

We develop a (nearly) unbiased particle filtering algorithm for a specific class of continuous-time state-space models, such that (a) the latent process $X_t$ is a linear Gaussian diffusion; and (b) the observations arise from a Poisson process with intensity $λ(X_t)$. The likelihood of the posterior probability density function of the latent process includes an intractable path integral. Our algorithm relies on Poisson estimates which approximate unbiasedly this integral. We show how we can tune these Poisson estimates to ensure that, with large probability, all but a few of the estimates generated by the algorithm are positive. Then replacing the negative estimates by zero leads to a much smaller bias than what would obtain through discretisation. We quantify the probability of negative estimates for certain special cases and show that our particle filter is effectively unbiased. We apply our method to a challenging 3D single molecule tracking example with a Born and Wolf observation model.

stat.CO↗

arXiv2022

Improved Gibbs samplers for Cosmic Microwave Background power spectrum estimation

We study different variants of the Gibbs sampler algorithm from the perspective of their applicability to the estimation of power spectra of the cosmic microwave background (CMB) anisotropies. These include approaches studied earlier in the CMB literature as well as new ones which are proposed in this work. We demonstrate all these variants on full and cut sky simulations and compare their performance, assessing both their computational and statistical efficiency. For this we employ a consistent comparison metric, an effective sample size (ESS) per second, commonly used in this context in the statistical literature. We show that one of the proposed approaches, referred to as Centered overrelax, which capitalizes on additional, auxiliary variables to minimize computational time needed per sample, and uses overrelaxation to decorrelate subsequent samples, performs better than the standard Gibbs sampler by a factor between one and two orders of magnitude in the nearly full-sky, satellite-like cases. It therefore potentially provides an interesting alternative to the currently favored approaches.

astro-ph.CO↗

arXiv2022

De-Sequentialized Monte Carlo: a parallel-in-time particle smoother

Particle smoothers are SMC (Sequential Monte Carlo) algorithms designed to approximate the joint distribution of the states given observations from a state-space model. We propose dSMC (de-Sequentialized Monte Carlo), a new particle smoother that is able to process $T$ observations in $\mathcal{O}(\log T)$ time on parallel architecture. This compares favourably with standard particle smoothers, the complexity of which is linear in $T$. We derive $\mathcal{L}_p$ convergence results for dSMC, with an explicit upper bound, polynomial in $T$. We then discuss how to reduce the variance of the smoothing estimates computed by dSMC by (i) designing good proposal distributions for sampling the particles at the initialization of the algorithm, as well as by (ii) using lazy resampling to increase the number of particles used in dSMC. Finally, we design a particle Gibbs sampler based on dSMC, which is able to perform parameter inference in a state-space model at a $\mathcal{O}(\log(T))$ cost on parallel hardware.

stat.CO↗

arXiv2021

Waste-free Sequential Monte Carlo

A standard way to move particles in a SMC sampler is to apply several steps of a MCMC (Markov chain Monte Carlo) kernel. Unfortunately, it is not clear how many steps need to be performed for optimal performance. In addition, the output of the intermediate steps are discarded and thus wasted somehow. We propose a new, waste-free SMC algorithm which uses the outputs of all these intermediate MCMC steps as particles. We establish that its output is consistent and asymptotically normal. We use the expression of the asymptotic variance to develop various insights on how to implement the algorithm in practice. We develop in particular a method to estimate, from a single run of the algorithm, the asymptotic variance of any particle estimate. We show empirically, through a range of numerical examples, that waste-free SMC tends to outperform standard SMC samplers, and especially so in situations where the mixing of the considered MCMC kernels decreases across iterations (as in tempering or rare event problems).

stat.CO↗

arXiv2021

Fast compression of MCMC output

We propose cube thinning, a novel method for compressing the output of a MCMC (Markov chain Monte Carlo) algorithm when control variates are available. It amounts to resampling the initial MCMC sample (according to weights derived from control variates), while imposing equality constraints on averages of these control variates, using the cube method of \cite{Deville2004}. Its main advantage is that its CPU cost is linear in $N$, the original sample size, and is constant in $M$, the required size for the compressed sample. This compares favourably to Stein thinning \citep{Riabiz2020}, which has complexity $\mathcal{O}(NM^2)$, and which requires the availability of the gradient of the target log-density (which automatically implies the availability of control variates). Our numerical experiments suggest that cube thinning is also competitive in terms of statistical error.

stat.CO↗

arXiv2020

Metropolis-Hastings with Averaged Acceptance Ratios

Markov chain Monte Carlo (MCMC) methods to sample from a probability distribution $π$ defined on a space $(Θ,\mathcal{T})$ consist of the simulation of realisations of Markov chains $\{θ_{n},n\geq1\}$ of invariant distribution $π$ and such that the distribution of $θ_{i}$ converges to $π$ as $i\rightarrow\infty$. In practice one is typically interested in the computation of expectations of functions, say $f$, with respect to $π$ and it is also required that averages $M^{-1}\sum_{n=1}^{M}f(θ_{n})$ converge to the expectation of interest. The iterative nature of MCMC makes it difficult to develop generic methods to take advantage of parallel computing environments when interested in reducing time to convergence. While numerous approaches have been proposed to reduce the variance of ergodic averages, including averaging over independent realisations of $\{θ_{n},n\geq1\}$ simulated on several computers, techniques to reduce the "burn-in" of MCMC are scarce. In this paper we explore a simple and generic approach to improve convergence to equilibrium of existing algorithms which rely on the Metropolis-Hastings (MH) update, the main building block of MCMC. The main idea is to use averages of the acceptance ratio w.r.t. multiple realisations of random variables involved, while preserving $π$ as invariant distribution. The methodology requires limited change to existing code, is naturally suited to parallel computing and is shown on our examples to provide substantial performance improvements both in terms of convergence to equilibrium and variance of ergodic averages. In some scenarios gains are observed even on a serial machine.

stat.CO↗

arXiv2020

Adaptive Tuning Of Hamiltonian Monte Carlo Within Sequential Monte Carlo

Sequential Monte Carlo (SMC) samplers form an attractive alternative to MCMC for Bayesian computation. However, their performance depends strongly on the Markov kernels used to rejuvenate particles. We discuss how to calibrate automatically (using the current particles) Hamiltonian Monte Carlo kernels within SMC. To do so, we build upon the adaptive SMC approach of Fearnhead and Taylor (2013), and we also suggest alternative methods. We illustrate the advantages of using HMC kernels within an SMC sampler via an extensive numerical study.

stat.CO↗

arXiv2020

Negative association, ordering and convergence of resampling methods

We study convergence and convergence rates for resampling schemes. Our first main result is a general consistency theorem based on the notion of negative association, which is applied to establish the almost-sure weak convergence of measures output from Kitagawa's (1996) stratified resampling method. Carpenter et al's (1999) systematic resampling method is similar in structure but can fail to converge depending on the order of the input samples. We introduce a new resampling algorithm based on a stochastic rounding technique of Srinivasan (2001), which shares some attractive properties of systematic resampling, but which exhibits negative association and therefore converges irrespective of the order of the input samples. We confirm a conjecture made by Kitagawa (1996) that ordering input samples by their states in $\mathbb{R}$ yields a faster rate of convergence; we establish that when particles are ordered using the Hilbert curve in $\mathbb{R}^d$, the variance of the resampling error is ${\scriptscriptstyle\mathcal{O}}(N^{-(1+1/d)})$ under mild conditions, where $N$ is the number of particles. We use these results to establish asymptotic properties of particle algorithms based on resampling schemes that differ from multinomial resampling.

stat.CO↗

arXiv2018

Improving approximate Bayesian computation via quasi-Monte Carlo

ABC (approximate Bayesian computation) is a general approach for dealing with models with an intractable likelihood. In this work, we derive ABC algorithms based on QMC (quasi- Monte Carlo) sequences. We show that the resulting ABC estimates have a lower variance than their Monte Carlo counter-parts. We also develop QMC variants of sequential ABC algorithms, which progressively adapt the proposal distribution and the acceptance threshold. We illustrate our QMC approach through several examples taken from the ABC literature. Keywords: Approximate Bayesian computation, Likelihood-free inference, Quasi Monte Carlo, Randomized Quasi-Monte Carlo, Adaptive importance sampling

stat.CO↗

arXiv2018

On the utility of Metropolis-Hastings with asymmetric acceptance ratio

The Metropolis-Hastings algorithm allows one to sample asymptotically from any probability distribution $π$. There has been recently much work devoted to the development of variants of the MH update which can handle scenarios where such an evaluation is impossible, and yet are guaranteed to sample from $π$ asymptotically. The most popular approach to have emerged is arguably the pseudo-marginal MH algorithm which substitutes an unbiased estimate of an unnormalised version of $π$ for $π$. Alternative pseudo-marginal algorithms relying instead on unbiased estimates of the MH acceptance ratio have also been proposed. These algorithms can have better properties than standard PM algorithms. Convergence properties of both classes of algorithms are known to depend on the variability of the estimators involved and reduced variability is guaranteed to decrease the asymptotic variance of ergodic averages and will shorten the burn-in period, or convergence to equilibrium, in most scenarios of interest. A simple approach to reduce variability, amenable to parallel computations, consists of averaging independent estimators. However, while averaging estimators of $π$ in a pseudo-marginal algorithm retains the guarantee of sampling from $π$ asymptotically, naive averaging of acceptance ratio estimates breaks detailed balance, leading to incorrect results. We propose an original methodology which allows for a correct implementation of this idea. We establish theoretical properties which parallel those available for standard PM algorithms and discussed above. We demonstrate the interest of the approach on various inference problems. In particular we show that convergence to equilibrium can be significantly shortened, therefore offering the possibility to reduce a user's waiting time in a generic fashion when a parallel computing architecture is available.

stat.CO↗

arXiv2018

Noise contrastive estimation: asymptotics, comparison with MC-MLE

A statistical model is said to be un-normalised when its likelihood function involves an intractable normalising constant. Two popular methods for parameter inference for these models are MC-MLE (Monte Carlo maximum likelihood estimation), and NCE (noise contrastive estimation); both methods rely on simulating artificial data-points to approximate the normalising constant. While the asymptotics of MC-MLE have been established under general hypotheses (Geyer, 1994), this is not so for NCE. We establish consistency and asymptotic normality of NCE estimators under mild assumptions. We compare NCE and MC-MLE under several asymptotic regimes. In particular, we show that, when m goes to infinity while n is fixed (m and n being respectively the number of artificial data-points, and actual data-points), the two estimators are asymptotically equivalent. Conversely, we prove that, when the artificial data-points are IID, and when n goes to infinity while m/n converges to a positive constant, the asymptotic variance of a NCE estimator is always smaller than the asymptotic variance of the corresponding MC-MLE estimator. We illustrate the variance reduction brought by NCE through a numerical study.

math.ST↗

arXiv2017

Sequential quasi-Monte Carlo: Introduction for Non-Experts, Dimension Reduction, Application to Partly Observed Diffusion Processes

SMC (Sequential Monte Carlo) is a class of Monte Carlo algorithms for filtering and related sequential problems. Gerber and Chopin (2015) introduced SQMC (Sequential quasi-Monte Carlo), a QMC version of SMC. This paper has two objectives: (a) to introduce Sequential Monte Carlo to the QMC community, whose members are usually less familiar with state-space models and particle filtering; (b) to extend SQMC to the filtering of continuous-time state-space models, where the latent process is a diffusion. A recurring point in the paper will be the notion of dimension reduction, that is how to implement SQMC in such a way that it provides good performance despite the high dimension of the problem.

stat.CO↗

arXiv2017

Modelling dependency completion in sentence comprehension as a Bayesian hierarchical mixture process: A case study involving Chinese relative clauses

We present a case-study demonstrating the usefulness of Bayesian hierarchical mixture modelling for investigating cognitive processes. In sentence comprehension, it is widely assumed that the distance between linguistic co-dependents affects the latency of dependency resolution: the longer the distance, the longer the retrieval time (the distance-based account). An alternative theory, direct-access, assumes that retrieval times are a mixture of two distributions: one distribution represents successful retrievals (these are independent of dependency distance) and the other represents an initial failure to retrieve the correct dependent, followed by a reanalysis that leads to successful retrieval. We implement both models as Bayesian hierarchical models and show that the direct-access model explains Chinese relative clause reading time data better than the distance account.

stat.AP↗

arXiv2016

Control functionals for Monte Carlo integration

A non-parametric extension of control variates is presented. These leverage gradient information on the sampling density to achieve substantial variance reduction. It is not required that the sampling density be normalised. The novel contribution of this work is based on two important insights; (i) a trade-off between random sampling and deterministic approximation and (ii) a new gradient-based function space derived from Stein's identity. Unlike classical control variates, our estimators achieve super-root-$n$ convergence, often requiring orders of magnitude fewer simulations to achieve a fixed level of precision. Theoretical and empirical results are presented, the latter focusing on integration problems arising in hierarchical models and models based on non-linear ordinary differential equations.

stat.ME↗

arXiv2015

Divide and conquer in ABC: Expectation-Progagation algorithms for likelihood-free inference

ABC algorithms are notoriously expensive in computing time, as they require simulating many complete artificial datasets from the model. We advocate in this paper a "divide and conquer" approach to ABC, where we split the likelihood into n factors, and combine in some way n "local" ABC approximations of each factor. This has two advantages: (a) such an approach is typically much faster than standard ABC and (b) it makes it possible to use local summary statistics (i.e. summary statistics that depend only on the data-points that correspond to a single factor), rather than global summary statistics (that depend on the complete dataset). This greatly alleviates the bias introduced by summary statistics, and even removes it entirely in situations where local summary statistics are simply the identity function. We focus on EP (Expectation-Propagation), a convenient and powerful way to combine n local approximations into a global approximation. Compared to the EP- ABC approach of Barthelmé and Chopin (2014), we present two variations, one based on the parallel EP algorithm of Cseke and Heskes (2011), which has the advantage of being implementable on a parallel architecture, and one version which bridges the gap between standard EP and parallel EP. We illustrate our approach with an expensive application of ABC, namely inference on spatial extremes.

stat.CO↗