Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Bayesian Inference”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,117 records · Page 62

Measuring the Viewing Angle of GW170817 with Electromagnetic and Gravitational Waves

The joint detection of gravitational waves (GWs) and electromagnetic (EM) radiation from the binary neutron star merger GW170817 ushered in a new era of multi-messenger astronomy. Joint GW–EM observations can be used to measure the parameters of the binary with better precision than either observation alone. Here, we use joint GW–EM observations to measure the viewing angle of GW170817, the angle between the binary's angular momentum and the line of sight. We combine a direct measurement of the distance to the host galaxy of GW170817 (NGC 4993) of 40.7 ± 2.36 Mpc with the Laser Interferometer Gravitational-wave Observatory (LIGO)/Virgo GW data and find that the viewing angle is ${32}_{-13}^{+10}\,\pm 1.7$ degrees (90% confidence, statistical, and systematic errors). We place a conservative lower limit on the viewing angle of ≥13°, which is robust to the choice of prior. This measurement provides a constraint on models of the prompt γ-ray and radio/X-ray afterglow emission associated with the merger; for example, it is consistent with the off-axis viewing angle inferred for a structured jet model. Lastly, we provide for the first time the full posterior samples from Bayesian parameter estimation of LIGO/Virgo data to enable further analysis by the community.

79 ASTRONOMY AND ASTROPHYSICS↗

A Review of Bayesian Networks for Spatial Data

We report Bayesian networks are a popular class of multivariate probabilistic models as they allow for the translation of prior beliefs about conditional dependencies between variables to be easily encoded into their model structure. Due to their widespread usage, they are often applied to spatial data for inferring properties of the systems under study and also generating predictions for how these systems may behave in the future. We review published research on methodologies for representing spatial data with Bayesian networks and also summarize the application areas for which Bayesian networks are employed in the modeling of spatial data. We find that a wide variety of perspectives are taken, including a GIS-centric focus on efficiently generating geospatial predictions, a statistical focus on rigorously constructing graphical models controlling for spatial correlation, as well as a range of problem-specific heuristics for mitigating the effects of spatial correlation and dependency arising in spatial data analysis. Special attention is also paid to potential future directions for integration of Bayesian networks with spatial processes.

97 MATHEMATICS AND COMPUTING↗

Bayesian Proteoform Modeling Improves Protein Quantification of Global Proteomic Measurements

As the capability of mass spectrometry-based proteomics has matured, tens of thousands of peptides can be measured simultaneously, which has the benefit of offering a systems view of protein expression. However, a major challenge is that with an increase in throughput, protein quantification estimation from the native measured peptides has become a computational task. A limitation to existing computationally-driven protein quantification methods is that most ignore protein variation, such as alternate splicing of the RNA transcript and post-translational modifications or other possible proteoforms, which will affect a significant fraction of the proteome. The consequence of this assumption is that statistical inference at the protein level, and consequently downstream analyses, such as network and pathway modeling, have only limited power for biomarker discovery. Here, we describe a Bayesian model (BP-Quant) that uses statistically derived peptides signatures to identify peptides that are outside the dominant pattern, or the existence of multiple over-expressed patterns to improve relative protein abundance estimates. It is a research-driven approach that utilizes the objectives of the experiment, defined in the context of a standard statistical hypothesis, to identify a set of peptides exhibiting similar statistical behavior relating to a protein. This approach infers that changes in relative protein abundance can be used as a surrogate for changes in function, without necessarily taking into account the effect of differential post-translational modifications, processing, or splicing in altering protein function. We verify the approach using a dilution study from mouse plasma samples and demonstrate that BP-Quant achieves similar accuracy as the current state-of-the-art methods at proteoform identification with significantly better specificity. BP-Quant is available as a MatLab ® and R packages at https://github.com/PNNL-Comp-Mass-Spec/BP-Quant.

Protein Quantitation, proteomics↗

Inverse prediction of PuO2 processing conditions using Bayesian seemingly unrelated regression with functional data

Over the past decade, a variety of innovative methodologies have been developed to better characterize the relationships between processing conditions and the physical, morphological, and chemical features of special nuclear material (SNM). Different processing conditions generate SNM products with different features, which are known as “signatures” because they are indicative of the processing conditions used to produce the material. These signatures can potentially allow a forensic analyst to determine which processes were used to produce the SNM and make inferences about where the material originated. This article investigates a statistical technique for relating processing conditions to the morphological features of PuO 2 particles. We develop a Bayesian implementation of seemingly unrelated regression (SUR) to inverse-predict unknown PuO 2 processing conditions from known PuO 2 features. Model results from simulated data demonstrate the usefulness of the technique. Applied to empirical data from a bench-scale experiment specifically designed with inverse prediction in mind, our model successfully predicts nitric acid concentration, while results for Pu concentration and precipitation temperature were equivalent to a simple mean model. Our technique compliments other recent methodologies developed for forensic analysis of nuclear material and can be generalized across the field of chemometrics for application to other materials.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Variational Bayesian Blind Image Deconvolution: A review

In this study, we provide a review of the recent literature on Bayesian Blind Image Deconvolution (BID) methods. We believe that two events have marked the recent history of BID: the predominance of Variational Bayes (VB) inference as a tool to solve BID problems and the increasing interest of the computer vision community in solving BID problems. VB inference in combination with recent image models like the ones based on Super Gaussian (SG) and Scale Mixture of Gaussians (SMG) representations have led to the use of very general and powerful tools to provide clear images from blurry observations. In the provided review emphasis is paid on VB inference and the use of SG and SMG models with coverage of recent advances in sampling methods. We also provide examples of current state of the art BID methods and discuss problems that very likely will mark the near future of BID.

97 MATHEMATICS AND COMPUTING↗

Disentangling the Black Hole Mass Spectrum with Photometric Microlensing Surveys

Abstract From the formation mechanisms of stars and compact objects to nuclear physics, modern astronomy frequently leverages surveys to understand populations of objects to answer fundamental questions. The population of dark and isolated compact objects in the Galaxy contains critical information related to many of these topics, but is only practically accessible via gravitational microlensing. However, photometric microlensing observables are degenerate for different types of lenses, and one can seldom classify an event as involving either a compact object or stellar lens on its own. To address this difficulty, we apply a Bayesian framework that treats lens type probabilistically and jointly with a lens population model. This method allows lens population characteristics to be inferred despite intrinsic uncertainty in the lens class of any single event. We investigate this method’s effectiveness on a simulated ground-based photometric survey in the context of characterizing a hypothetical population of primordial black holes (PBHs) with an average mass of 30 M ⊙ . On simulated data, our method outperforms current black hole (BH) lens identification pipelines and characterizes different subpopulations of lenses while jointly constraining the PBH contribution to dark matter to ≈25%. Key to robust inference, our method can marginalize over population model uncertainty. We find the lower mass cutoff for stellar origin BHs, a key observable in understanding the BH mass gap, particularly difficult to infer in our simulations. This work lays the foundation for cutting-edge PBH abundance constraints to be extracted from current photometric microlensing surveys.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

A Robust Estimate of Continental-Scale Terrestrial Carbon Sinks Using GOSAT XCO 2 Retrievals

Satellite XCO 2 retrievals could improve the estimates of surface carbon fluxes, but it remains unknown on what scales these estimates are robust. Here, we use the time-dependent Bayesian synthesis top-down method and prior net ecosystem exchanges (NEEs) from 12 terrestrial biosphere models (TBMs) to infer the monthly carbon fluxes of 51 land regions with constraints by GOSAT XCO 2 retrievals. We find that the uncertainty (standard deviation of 12 TBMs) reduction rates (uncertainty reduction rate (URR)) decrease significantly at decreasing spatial scales. On the continental-scale, the mean URR is about 57%, and the annual and seasonal cycle estimates of NEE are rather robust. The evaluation shows that the posterior CO 2 concentrations are significantly improved at the continental scale. Our study suggests that the GOSAT XCO 2 can only promise a robust continental-scale NEE estimate and improving the XCO 2 accuracy is an effective way to achieve robust estimates on smaller scales under current spatial coverage.

58 GEOSCIENCES↗

MODEL SELECTION FOR SPECTROPOLARIMETRIC INVERSIONS

Inferring magnetic and thermodynamic information from spectropolarimetric observations relies on the assumption of a parameterized model atmosphere whose parameters are tuned by comparison with observations. Often, the choice of the underlying atmospheric model is based on subjective reasons. In other cases, complex models are chosen based on objective reasons (for instance, the necessity to explain asymmetries in the Stokes profiles) but it is not clear what degree of complexity is needed. The lack of an objective way of comparing models has, sometimes, led to opposing views of the solar magnetism because the inferred physical scenarios are essentially different. We present the first quantitative model comparison based on the computation of the Bayesian evidence ratios for spectropolarimetric observations. Our results show that there is not a single model appropriate for all profiles simultaneously. Data with moderate signal-to-noise ratios (S/Ns) favor models without gradients along the line of sight. If the observations show clear circular and linear polarization signals above the noise level, models with gradients along the line are preferred. As a general rule, observations with large S/Ns favor more complex models. We demonstrate that the evidence ratios correlate well with simple proxies. Therefore, we propose to calculate these proxies when carrying out standard least-squares inversions to allow for model comparison in the future.

79 ASTRONOMY AND ASTROPHYSICS↗

Forward variable selection enables fast and accurate dynamic system identification with Karhunen-Loève decomposed Gaussian processes

A promising approach for scalable Gaussian processes (GPs) is the Karhunen-Loève (KL) decomposition, in which the GP kernel is represented by a set of basis functions which are the eigenfunctions of the kernel operator. Such decomposed kernels have the potential to be very fast, and do not depend on the selection of a reduced set of inducing points. However KL decompositions lead to high dimensionality, and variable selection thus becomes paramount. This paper reports a new method of forward variable selection, enabled by the ordered nature of the basis functions in the KL expansion of the Bayesian Smoothing Spline ANOVA kernel (BSS-ANOVA), coupled with fast Gibbs sampling in a fully Bayesian approach. It quickly and effectively limits the number of terms, yielding a method with competitive accuracies, training and inference times for tabular datasets of low feature set dimensionality. Theoretical computational complexities are O ( N P 2 ) in training and O ( P ) per point in inference, where N is the number of instances and P the number of expansion terms. The inference speed and accuracy makes the method especially useful for dynamic systems identification, by modeling the dynamics in the tangent space as a static problem, then integrating the learned dynamics using a high-order scheme. The methods are demonstrated on two dynamic datasets: a ‘Susceptible, Infected, Recovered’ (SIR) toy problem, along with the experimental ‘Cascaded Tanks’ benchmark dataset. Comparisons on the static prediction of time derivatives are made with a random forest (RF), a residual neural network (ResNet), and the Orthogonal Additive Kernel (OAK) inducing points scalable GP, while for the timeseries prediction comparisons are made with LSTM and GRU recurrent neural networks (RNNs) along with the SINDy package.

Hayes, Kyle↗

"Sintering" Models and In-Situ Experiments: Data Assimilation for Microstructure Prediction in SLS Additive Manufacturing of Nylon Components

Selective laser sintering methods are workhorses for additively manufacturing polymer-based components. The ease of rapid prototyping also means it is easy to produce illicit components. It is necessary to have a data-calibrated in-situ physical model of the build process in order to predict expected and defective microstructure characteristics that inform component provenance. Toward this end, sintering models are calibrated and characteristics such as component defects are explored. This is accomplished by assimilating multiple data streams, imaging analysis, and computational model predictions in an adaptive Bayesian parameter estimation algorithm. From these data sources, along with a phase-field model, bulk porosity distributions are inferred. Model parameters are constrained to physically-relevant search directions by sensitivity analysis, and then matched to predictions using adaptive sampling. Using this feedback loop, data-constrained estimates of sintering model parameters along with uncertainty bounds are obtained.

3D printing, Additive Manufacturing, polymer, sint↗

A Deep Generative Deconvolutional Image Model

A deep generative model is developed for representation and analysis of images, based on a hierarchical convolutional dictionary-learning framework. Stochastic unpooling is employed to link consecutive layers in the model, yielding top-down image generation. A Bayesian support vector machine is linked to the top-layer features, yielding max-margin discrimination. Deep deconvolutional inference is employed when testing, to infer the latent features, and the top-layer features are connected with the max-margin classifier for discrimination tasks. The model is efficiently trained using a Monte Carlo expectation-maximization (MCEM) algorithm; the algorithm is implemented on graphical processor units (GPU) to enable large-scale learning, and fast testing. Excellent results are obtained on several benchmark datasets, including ImageNet, demonstrating that the proposed model achieves results that are highly competitive with similarly sized convolutional neural networks.

convolutional neural network↗

THE MEASUREMENT, TREATMENT, AND IMPACT OF SPECTRAL COVARIANCE AND BAYESIAN PRIORS IN INTEGRAL -FIELD SPECTROSCOPY OF EXOPLANETS

The recovery of an exoplanet’s atmospheric parameters from its spectrum requires accurate knowledge of the spectral errors and covariances. Unfortunately, the complex image processing used in high-contrast integral-field spectrograph (IFS) observations generally produces spectral covariances that are poorly understood and often ignored. In this work, we show how to measure the spectral errors and covariances and include them self-consistently in parameter retrievals. By combining model exoplanet spectra with a realistic noise model generated from the Gemini Planet Imager (GPI) early science data, we show that ignoring spectral covariance in high-contrast IFS data can both bias inferred parameters and lead to unreliable confidence regions on those parameters. This problem is made worse by the common practice of scaling the χ {sup 2} per degree of freedom to unity; the input parameters then fall outside the 95% confidence regions in as many as ∼80% of noise realizations. The biases we observe can approach the typical levels of precision achieved in high-contrast spectroscopy. Accounting for realistic priors in fully Bayesian retrievals can also have a significant impact on the inferred parameters. Plausible priors on effective temperature and surface gravity can vary by an order of magnitude across the confidence regions appropriate for objects with weak age constraints; priors for objects with good age constraints are dominated by modeling uncertainties. Our methods are directly applicable to existing high-contrast IFSs including GPI and SPHERE, as well as upcoming instruments like CHARIS and, ultimately, WFIRST-AFTA.

79 ASTRONOMY AND ASTROPHYSICS↗

Scalable optimal Bayesian classification of single-cell trajectories under regulatory model uncertainty

Single-cell gene expression measurements offer opportunities in deriving mechanistic understanding of complex diseases, including cancer. However, due to the complex regulatory machinery of the cell, gene regulatory network (GRN) model inference based on such data still manifests significant uncertainty. The goal of this paper is to develop optimal classification of single-cell trajectories accounting for potential model uncertainty. Partially-observed Boolean dynamical systems (POBDS) are used for modeling gene regulatory networks observed through noisy gene-expression data. We derive the exact optimal Bayesian classifier (OBC) for binary classification of single-cell trajectories. The application of the OBC becomes impractical for large GRNs, due to computational and memory requirements. To address this, we introduce a particle-based single-cell classification method that is highly scalable for large GRNs with much lower complexity than the optimal solution. The performance of the proposed particle-based method is demonstrated through numerical experiments using a POBDS model of the well-known T-cell large granular lymphocyte (T-LGL) leukemia network with noisy time-series gene-expression data.

59 BASIC BIOLOGICAL SCIENCES↗

Simultaneous inference of the compressibility and inelastic response of tantalum under extreme loading

We study the deformation of tantalum under extreme loading conditions. Experimental velocity data are drawn from both ramp loading experiments on Sandia’s Z-machine and gas gun compression experiments. The drive conditions enable the study of materials under pressures greater than 100 GPa. We provide a detailed forward model of the experiments including a model of the magnetic drive for the Z-machine. Utilizing these experiments, we simultaneously infer several different types of physically motivated parameters describing equation of state, plasticity, and anelasticity via the computational device of Bayesian model calibration. Characteristics of the resulting calculated posterior distributions illustrate relationships among the parameters of interest via the degree of cross correlation. The calibrated velocity traces display good agreement with the experiments up to experimental uncertainty as well as improvement over previous calibrations. Examining the Z-shots and gun-shots together and separately reveals a trade-off between accuracy and transferability across different experimental conditions. Implications for model calibration, limitations from model form, and suggestions for improvements are discussed.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Correlations between the Neutron Star Mass–Radius Relation and the Equation of State of Dense Matter

We develop an analytic method of inverting the Tolman–Oppenheimer–Volkoff relations to high accuracy. In principle, a specified energy density–pressure relation gives a unique mass–radius (M–R) relation and vice versa. Our method is developed from the strong correlations that are shown to exist between the neutron star mass–radius curve and the equation of state (EOS) or pressure–energy density relation. Selecting points that have masses equal to fixed fractions of the maximum mass, we find a semi-universal power-law relation between the central energy densities, pressures, sound speeds, chemical potentials, and number densities of those stars, with the maximum mass and the radii of one or more fractional maximum mass points. Rms fitting accuracies, for EOSs without large first-order phase transitions, are typically 0.5% for all quantities at all mass points. The method also works well, although less accurately, in reconstructing the EOS of hybrid stars with first-order phase transitions. These results permit, in effect, an analytic method of inverting an arbitrary M–R curve to yield its underlying EOS. We discuss applications of this inversion technique to the inference of the dense matter EOS from measurements of neutron star masses and radii as a possible alternative to traditional Bayesian approaches.

Bayesian statistics↗

Multimodel ensemble predictions of precipitation using bayesian neural networks

Multimodel ensembling improves predictions and considers model uncertainties. In this study, we present a Bayesian Neural Network (BNN) ensemble approach for large-scale precipitation predictions based on a set of climate models. BNN infers spatiotemporally varying model weights and biases through the calibration against observations. This ensemble scheme of BNN sufficiently leverages individual model skill for accurate predictions as well as provides interpretability about which models contribute more to the ensemble prediction at which locations and times to inform model development. Additionally, BNN accurately quantifies epistemic uncertainty to avoid overconfident projections. We demonstrate BNN’s superior prediction performance to three state-of-the-art ensemble approaches and discuss its interpretability and uncertainty quantification.

Fan, Patrick↗

Multimodel Ensemble Predictions of Precipitation using Bayesian Neural Networks

Multimodel ensembling improves predictions and considers model uncertainties. In this study, we present a Bayesian Neural Network (BNN) ensemble approach for large-scale precipitation predictions based on a set of climate models. BNN infers spatiotemporally varying model weights and biases through the calibration against observations. This ensemble scheme of BNN sufficiently leverages individual model skill for accurate predictions as well as provides interpretability about which models contribute more to the ensemble prediction at which locations and times to inform model development. Additionally, BNN accurately quantifies epistemic uncertainty to avoid overconfident projections. We demonstrate BNN’s superior prediction performance to three state-of-the-art ensemble approaches and discuss its interpretability and uncertainty quantification.

Fan, Patrick↗

Scalable and efficient learning from crowds with Gaussian processes

Over the last few years, multiply-annotated data has become a very popular source of information. Online platforms such as Amazon Mechanical Turk have revolutionized the labelling process needed for any classification task, sharing the effort between a number of annotators (instead of the classical single expert). This crowdsourcing approach has introduced new challenging problems, such as handling disagreements on the annotated samples or combining the unknown expertise of the annotators. Probabilistic methods, such as Gaussian Processes (GP), have proven successful to model this new crowdsourcing scenario. However, GPs do not scale up well with the training set size, which makes them prohibitive for medium-to-large datasets (beyond 10K training instances). This constitutes a serious limitation for current real-world applications. In this work, we introduce two scalable and efficient GP-based crowdsourcing methods that allow for processing previously-prohibitive datasets. The first one is an efficient and fast approximation to GP with squared exponential (SE) kernel. The second allows for learning a more flexible kernel at the expense of a heavier training (but still scalable to large datasets). Since the latter is not a GP-SE approximation, it can be also considered as a whole new scalable and efficient crowdsourcing method, useful for any dataset size. Both methods use Fourier features and variational inference, can predict the class of new samples, and estimate the expertise of the involved annotators. A complete experimentation compares them with state-of-the-art probabilistic approaches in synthetic and real crowdsourcing datasets of different sizes. Finally, they stand out as the best performing approach for large scale problems. Moreover, the second method is competitive with the current state-of-the-art for small datasets.

96 KNOWLEDGE MANAGEMENT AND PRESERVATION↗