A Bayesian-Deep Learning Model for Estimating COVID-19 Evolution in Spain

This work proposes a semi-parametric approach to estimate the evolution of COVID-19 (SARS-CoV-2) in Spain. Considering the sequences of 14-day cumulative incidence of all Spanish regions, it combines modern Deep Learning (DL) techniques for analyzing sequences with the usual Bayesian Poisson-Gamma model for counts. The DL model provides a suitable description of the observed time series of counts, but it cannot give a reliable uncertainty quantification. The role of expert elicitation of the expected number of counts and its reliability is DL predictions’ role in the proposed modelling approach. Finally, the posterior predictive distribution of counts is obtained in a standard Bayesian analysis using the well known Poisson-Gamma model. The model allows to predict the future evolution of the sequences on all regions or estimates the consequences of eventual scenarios.

The hubble tension as a hint of leptogenesis and neutrino mass generation

The majoron, a neutrinophilic pseudo-Goldstone boson conventionally arising in the context of neutrino mass models, can damp neutrino free-streaming and inject additional energy density into neutrinos prior to recombination. The combination of these effects for an eV-scale mass majoron has been shown to ameliorate the outstanding $$H_0$$ H 0 tension, however only if one introduces additional dark radiation at the level of $$\Delta N_{\mathrm{eff}} \sim 0.5$$ Δ N eff ∼ 0.5 . We show here that models of low-scale leptogenesis can naturally source this dark radiation by generating a primordial population of majorons from the decays of GeV-scale sterile neutrinos in the early Universe. Using a posterior predictive distribution conditioned on Planck2018+BAO data, we show that the value of $$H_0$$ H 0 observed by the SH$$_0$$ 0 ES collaboration is expected to occur at the level of $$\sim 10\%$$ ∼ 10 % in the primordial majoron cosmology (to be compared with $$\sim 0.1\%$$ ∼ 0.1 % in the case of $$\Lambda $$ Λ CDM). This insight provides an intriguing connection between the neutrino mass mechanism, the baryon asymmetry of the Universe, and the discrepant measurements of $$H_0$$ H 0 .

Weighted Dirichlet Process Mixture Models to Accommodate Complex Sample Designs for Linear and Quantile Regression

Standard randomization-based inference conditions on the data in the population and makes inference with respect to the repeating sampling properties of the sampling indicators. In some settings these estimators can be quite unstable; Bayesian model-based approaches focus on the posterior predictive distribution of population quantities, potentially providing a better balance between bias correction and efficiency. Previous work in this area has focused on estimation of means and linear and generalized linear regression parameters; these methods do not allow for a general estimation of distributional functions such as quantile or quantile regression parameters. Here we adapt an extended Dirichlet Process Mixture model that allows the DP prior to be a mixture of DP random basis measures that are a function of covariates. These models allow many mixture components when necessary to accommodate the sample design, but can shrink to few components for more efficient estimation when the data allow. We provide an application to the estimation of relationships between serum dioxin levels and age in the US population, either at the mean level (via linear regression) or across the dioxin distribution (via quantile regression) using the National Health and Nutrition Examination Survey.

Dark energy survey internal consistency tests of the joint cosmological probes analysis with posterior predictive distributions

ABSTRACT Beyond ΛCDM, physics or systematic errors may cause subsets of a cosmological data set to appear inconsistent when analysed assuming ΛCDM. We present an application of internal consistency tests to measurements from the Dark Energy Survey Year 1 (DES Y1) joint probes analysis. Our analysis relies on computing the posterior predictive distribution (PPD) for these data under the assumption of ΛCDM. We find that the DES Y1 data have an acceptable goodness of fit to ΛCDM, with a probability of finding a worse fit by random chance of p = 0.046. Using numerical PPD tests, supplemented by graphical checks, we show that most of the data vector appears completely consistent with expectations, although we observe a small tension between large- and small-scale measurements. A small part (roughly 1.5 per cent) of the data vector shows an unusually large departure from expectations; excluding this part of the data has negligible impact on cosmological constraints, but does significantly improve the p-value to 0.10. The methodology developed here will be applied to test the consistency of DES Year 3 joint probes data sets.

A Model Comparison Approach to Posterior Predictive Model Checks in Bayesian Confirmatory Factor Analysis

Posterior Predictive Model Checking (PPMC) is frequently used for model fit evaluation in Bayesian Confirmatory Factor Analysis (BCFA). In standard PPMC procedures, model misfit is quantified by the location of a ML-based estimate to the predictive distribution of a statistic for a model. When the ML-based point estimate is far away from the center of the density of the posterior predictive distribution, model fit is poor. One main critique of such standard PPMC procedures is the strong link to the ML-based point estimates of the observed data. Not included in this approach, however, is how variable the ML-based point estimates are and their use in general as the reference point for Bayesian analyses. We propose a new method of PPMC based on the Posterior Predictive distribution of Bayesian saturated model for BCFA models. The method uses the predictive distribution from parameters of the posterior distribution of the saturated model as reference to detect the local misfit of hypothesized models. The results of the simulation study suggest that the saturated model PPMC approach was an accurate method of determining local model misfit and could be used for model comparison. A real example is also provided in this study.

Binary Time Series Classification with Bayesian Convolutional Neural Networks When Monitoring for Marine Gas Discharges

The world’s oceans are under stress from climate change, acidification and other human activities, and the UN has declared 2021–2030 as the decade for marine science. To monitor the marine waters, with the purpose of detecting discharges of tracers from unknown locations, large areas will need to be covered with limited resources. To increase the detectability of marine gas seepage we propose a deep probabilistic learning algorithm, a Bayesian Convolutional Neural Network (BCNN), to classify time series of measurements. The BCNN will classify time series to belong to a leak/no-leak situation, including classification uncertainty. The latter is important for decision makers who must decide to initiate costly confirmation surveys and, hence, would like to avoid false positives. Results from a transport model are used for the learning process of the BCNN and the task is to distinguish the signal from a leak hidden within the natural variability. We show that the BCNN classifies time series arising from leaks with high accuracy and estimates its associated uncertainty. We combine the output of the BCNN model, the posterior predictive distribution, with a Bayesian decision rule showcasing how the framework can be used in practice to make optimal decisions based on a given cost function.

Identifying model violations under the multispecies coalescent model using P2C2M.SNAPP

Phylogenetic estimation under the multispecies coalescent model (MSCM) assumes all incongruence among loci is caused by incomplete lineage sorting. Therefore, applying the MSCM to datasets that contain incongruence that is caused by other processes, such as gene flow, can lead to biased phylogeny estimates. To identify possible bias when using the MSCM, we present P2C2M.SNAPP. P2C2M.SNAPP is an R package that identifies model violations using posterior predictive simulation. P2C2M.SNAPP uses the posterior distribution of species trees output by the software package SNAPP to simulate posterior predictive datasets under the MSCM, and then uses summary statistics to compare either the empirical data or the posterior distribution to the posterior predictive distribution to identify model violations. In simulation testing, P2C2M.SNAPP correctly classified up to 83% of datasets (depending on the summary statistic used) as to whether or not they violated the MSCM model. P2C2M.SNAPP represents a user-friendly way for researchers to perform posterior predictive model checks when using the popular SNAPP phylogenetic estimation program. It is freely available as an R package, along with additional program details and tutorials.