Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Variational Bayesian”

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 91 records · Page 5

Understanding the Scalability of Bayesian Network Inference using Clique Tree Growth Curves

Bayesian networks (BNs) are used to represent and efficiently compute with multi-variate probability distributions in a wide range of disciplines. One of the main approaches to perform computation in BNs is clique tree clustering and propagation. In this approach, BN computation consists of propagation in a clique tree compiled from a Bayesian network. There is a lack of understanding of how clique tree computation time, and BN computation time in more general, depends on variations in BN size and structure. On the one hand, complexity results tell us that many interesting BN queries are NP-hard or worse to answer, and it is not hard to find application BNs where the clique tree approach in practice cannot be used. On the other hand, it is well-known that tree-structured BNs can be used to answer probabilistic queries in polynomial time. In this article, we develop an approach to characterizing clique tree growth as a function of parameters that can be computed in polynomial time from BNs, specifically: (i) the ratio of the number of a BN's non-root nodes to the number of root nodes, or (ii) the expected number of moral edges in their moral graphs. Our approach is based on combining analytical and experimental results. Analytically, we partition the set of cliques in a clique tree into different sets, and introduce a growth curve for each set. For the special case of bipartite BNs, we consequently have two growth curves, a mixed clique growth curve and a root clique growth curve. In experiments, we systematically increase the degree of the root nodes in bipartite Bayesian networks, and find that root clique growth is well-approximated by Gompertz growth curves. It is believed that this research improves the understanding of the scaling behavior of clique tree clustering, provides a foundation for benchmarking and developing improved BN inference and machine learning algorithms, and presents an aid for analytical trade-off studies of clique tree clustering using growth curves.

Mengshoel, Ole Jakob↗

Image processing of multiphase images obtained via X-ray microtomography: A review

Easier access to X-ray microtomography (μCT) facilities has provided much new insight from high-resolution imaging for various problems in porous media research. Pore space analysis with respect to functional properties usually requires segmentation of the intensity data into different classes. Image segmentation is a nontrivial problem that may have a profound impact on all subsequent image analyses. This review deals with two issues that are neglected in most of the recent studies on image segmentation: (i) focus on multiclass segmentation and (ii) detailed descriptions as to why a specific method may fail together with strategies for preventing the failure by applying suitable image enhancement prior to segmentation. In this way, the presented algorithms become very robust and are less prone to operator bias. Three different test images are examined: a synthetic image with ground-truth information, a synchrotron image of precision beads with three different fluids residing in the pore space, and a μCT image of a soil sample containing macropores, rocks, organic matter, and the soil matrix. Image blur is identified as the major cause for poor segmentation results. Other impairments of the raw data like noise, ring artifacts, and intensity variation can be removed with current image enhancement methods. Bayesian Markov random field segmentation, watershed segmentation, and converging active contours are well suited for multiclass segmentation, yet with different success to correct for partial volume effects and conserve small image features simultaneously.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

A stochastic approach to quantifying the blur with uncertainty estimation for high-energy X-ray imaging systems

One of the primary causes of blur in a high-energy X-ray imaging system is the shape and extent of the radiation source, or ‘spot’. It is important to be able to quantify the size of the spot as it provides a lower bound on the recoverable resolution for a radiograph, and penumbral imaging methods – which involve the analysis of blur caused by a structured aperture – can be used to obtain the spot’s spatial profile. We present a Bayesian approach for estimating the spot shape that, unlike variational methods, is robust to the initial choice of parameters. The posterior is obtained from a normal likelihood, which was constructed from a weighted least squares approximation to a Poisson noise model, and prior assumptions that enforce both smoothness and non-negativity constraints. A Markov chain Monte Carlo algorithm is used to obtain samples from the target posterior, and the reconstruction and uncertainty estimates are the computed mean and variance of the samples, respectively. Lastly, synthetic data-sets are used to demonstrate accurate reconstruction, while real data taken with high-energy X-ray imaging systems are used to demonstrate applicability and feasibility.

97 MATHEMATICS AND COMPUTING↗

Adaptive Discovery and Mixed-Variable Optimization of Next Generation Synthesizable Microelectronic Materials

Design of new microelectronic materials is characterized by several challenges such as high-dimensionality of the atomic structure-composition variable space, formidable cost of directly using high-fidelity simulations for design optimization, dispersity in literature-reported similar materials and synthesis methods, complex physical mechanisms, and mixed qualitative and quantitative design variables that lead to a disjointed design space. Even though machine learning (ML) techniques have been employed to expedite materials innovation, existing methods treat ML and design optimization as two separate processes, failing to resolve the fundamental challenges associated with high dimensionality and mixed-variable complexity. We have developed a ML enhanced mixed-variable material design optimization framework to efficiently extract useful information from existing data in literature and physics-based simulations to guide the autonomous search for optimal materials. Our proposed framework is composed of four computational modules: (1) a natural language processing (NLP) based virtual screening module, (2) classification based concept exploration module, (3) a density functional theory (DFT)-based high-fidelity evaluation model, and (4) a novel latent-variable Gaussian process (LVGP) ML model for mixed-variable problems with uncertainty quantification, which seamlessly integrates with Bayesian Optimization (BO) and achieves superb efficiency through embedded physics-based dimension reduction. Our approach is demonstrated and validated using the testbed of functional materials exhibiting metal-insulation transitions (MITs), with the targeted reversible resistivity changes (∼10^5) near room temperature. At the end of the 30-month project, we have developed a series of new ML techniques using NLP, conditional variational autoencoders, active learning, latent-variable Gaussian processes, integrated with Bayesian optimization. Our project has resulted in new predicted MITs compounds and improved understanding of MITs microscopic mechanisms, which in turn will revolutionize microelectronics science to provide energy-saving solutions. Our research has improved both creativity and efficiency in transforming rare-event discoveries of new functional materials to persistent innovations. In addition to open-sourcing the online MIT database and the classification model, the LVGP open source code has been downloaded more than 15,000 times within two years. More than 40 MIT compounds have been identified and many have been pursued experimentally via collaborators. The research results are published in close to 20 collaborative papers in high-impact journals, such as Chem. Mater., Appl. Phys. Rev., Sci. Rep., among others of design space.

36 MATERIALS SCIENCE↗

A bayesian approach to estimate the size and structure of the broad-line region in active galactic nuclei using reverberation mapping data

This is the first paper in a series devoted to the systematic study of the size and structure of the broad-line region (BLR) in active galactic nuclei (AGNs) using reverberation mapping (RM) data. We employ a recently developed Bayesian approach that statistically describes the variability as a damped random walk process and delineates the BLR structure using a flexible disk geometry that can account for a variety of shapes, including disks, rings, shells, and spheres. We allow for the possibility that the line emission may respond non-linearly to the continuum, and we detrend the light curves when there is clear evidence for secular variation. We use a Markov Chain Monte Carlo implementation based on Bayesian statistics to recover the parameters and uncertainties for the BLR model. The corresponding transfer function is obtained self-consistently. We tentatively constrain the virial factor used to estimate black hole masses; more accurate determinations will have to await velocity-resolved RM data. Application of our method to RM data with Hβ monitoring for about 40 objects shows that the assumed BLR geometry can reproduce quite well the observed emission-line fluxes from the continuum light curves. We find that the Hβ BLR sizes obtained from our method are on average ∼20% larger than those derived from the traditional cross-correlation method. Nevertheless, we still find a tight BLR size-luminosity relation with a slope of α = 0.55 ± 0.03 and an intrinsic scatter of ∼0.18 dex. In particular, we demonstrate that our approach yields appropriate BLR sizes for some objects (such as Mrk 142 and PG 2130+099) where traditional methods previously encountered difficulties.

79 ASTRONOMY AND ASTROPHYSICS↗

A scalable variational method for estimating the latent infection-rate field of an outbreak

In this paper, we explore whether the infection-rate of a disease can serve as a robust monitoring variable in epidemiological surveillance algorithms. The infection-rate is dependent on population mixing patterns that do not vary erratically day-to-day; in contrast, daily case-counts used in contemporary surveillance algorithms are corrupted by reporting errors. The technical challenge lies in estimating the latent infection-rate from case-counts. Here we devise a Bayesian method to estimate the infection-rate across multiple adjoining areal units, and then use it, via an anomaly detector, to discern a change in epidemiological dynamics. We extend an existing model for estimating the infection-rate in an areal unit by incorporating a Markov random field model, so that we may estimate infection-rates across multiple areal units, while preserving spatial correlations observed in the epidemiological dynamics. To carry out the high-dimensional Bayesian inverse problem, we develop an implementation of mean-field variational inference specific to the infection model and integrate it with the random field model to incorporate correlations across counties. The method is tested on estimating the COVID-19 infection-rates across all 33 counties in New Mexico using data from the summer of 2020, and then employing them to detect the arrival of the Fall 2020 COVID-19 wave. We perform the detection using a temporal algorithm that is applied county-by-county. We also show how the infection-rate field can be used to cluster counties with similar epidemiological dynamics.

60 APPLIED LIFE SCIENCES↗

Is Bayesian inference "brittle"?

In a recent report, “Bayesian Brittleness: Why no Bayesian model is ‘good enough,’” (arXiv, 1304.6772v1) the authors prove rigorously that model variations that are arbitrarily small, in a particular technical sense, the posterior expectation of a function can achieve essentially any value that the function alone can achieve, so that Bayesian inference appears to have no robustness whatsoever. We explain this puzzling result, and show why it does not imply a breakdown of Bayesian inference. The explanation is that the models leading to the extreme results depend on the observed data.

97 MATHEMATICS AND COMPUTING↗

Modeling Radiation Detection Using a Mixed-Effects Statistical Model for Background Variation

In this paper, we consider radiation source localization in an urban environment as an inverse problem and use Bayesian parameter estimation techniques to infer the source location and intensity. Since parameter estimation generally requires a large number of model evaluations, we employ a simplified photon transport model, as in a 2017 paper by Stefanescu, et al., to reduce computational requirements. However, the model in the 2017 paper does not account for the physical reality that background radiation varies with position within an urban locale. We incorporate statistical mixed-effects modeling into the simplified transport equation from the Stefanescu, et al. paper to reflect individual variation among the background radiation measurements at multiple detector locations. With our new model, we use a Bayesian DRAM-within-Gibbs technique, which incorporates the Delayed Rejection Adaptive Metropolis (DRAM) algorithm, to estimate the mixed-effects parameters. We note that without adequate prior information, we were unable to simultaneously estimate the set of parameters corresponding to the source intensity, source location, and individual background radiation counts. By calibrating the background in the absence of a source, we obtained narrow prior distributions, which permitted us to jointly infer the source intensity and location and the background radiation distribution.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Avian collision risk models for wind energy impact assessments

With the increasing global development of wind energy, collision risk models (CRMs) are routinely used to assess the potential impacts of wind turbines on birds. We reviewed and compared the avian collision risk models currently available in the scientific literature, exploring aspects such as the calculation of a collision probability, inclusion of stationary components e.g. the tower, angle of approach and uncertainty. 10 models were cited in the literature and of these, all included a probability of collision of a single bird colliding with a wind turbine during passage through the rotor swept area, and the majority included a measure of the number of birds at risk. 7 out of the 10 models calculated the probability of birds colliding, whilst the remainder used a constant. We identified four approaches to calculate the probability of collision and these were used by others. 6 of the 10 models were deterministic and included the most frequently used models in the UK, with only 4 including variation or uncertainty in some way, the most recent using Bayesian methods. Despite their appeal, CRMs have their limitations and can be ‘data hungry’ as well as assuming much about bird movement and behaviour. As data become available, these assumptions should be tested to ensure that CRMs are functioning to adequately answer the questions posed by the wind energy sector. - Highlights: • We highlighted ten models available to assess avian collision risk. • Only 4 of the models included variability or uncertainty. • Collision risk models have limitations and can be ‘data hungry’. • It is vital that the most appropriate model is used for a given task.

54 ENVIRONMENTAL SCIENCES↗

A unifying Bayesian framework for merging X-ray diffraction data

Novel X-ray methods are transforming the study of the functional dynamics of biomolecules. Key to this revolution is detection of often subtle conformational changes from diffraction data. Diffraction data contain patterns of bright spots known as reflections. To compute the electron density of a molecule, the intensity of each reflection must be estimated, and redundant observations reduced to consensus intensities. Systematic effects, however, lead to the measurement of equivalent reflections on different scales, corrupting observation of changes in electron density. Here, we present a modern Bayesian solution to this problem, which uses deep learning and variational inference to simultaneously rescale and merge reflection observations. We successfully apply this method to monochromatic and polychromatic single-crystal diffraction data, as well as serial femtosecond crystallography data. We find that this approach is applicable to the analysis of many types of diffraction experiments, while accurately and sensitively detecting subtle dynamics and anomalous scattering.

59 BASIC BIOLOGICAL SCIENCES↗

How to distinguish between cloudy mini-Neptunes and water/volatile-dominated super-Earths

One of the most profound questions about the newly discovered class of low-density super-Earths is whether these exoplanets are predominately H{sub 2}-dominated mini-Neptunes or volatile-rich worlds with gas envelopes dominated by H{sub 2}O, CO{sub 2}, CO, CH{sub 4}, or N{sub 2}. Transit observations of the super-Earth GJ 1214b rule out cloud-free H{sub 2}-dominated scenarios, but are not able to determine whether the lack of deep spectral features is due to high-altitude clouds or the presence of a high mean molecular mass atmosphere. Here, we demonstrate that one can unambiguously distinguish between cloudy mini-Neptunes and volatile-dominated worlds based on wing steepness and relative depths of absorption features in moderate-resolution near-infrared transmission spectra (R ∼ 100). In a numerical retrieval study, we show for GJ 1214b that an unambiguous distinction between a cloudy H{sub 2}-dominated atmosphere and cloud-free H{sub 2}O atmosphere will be possible if the uncertainties in the spectral transit depth measurements can be reduced by a factor of ∼3 compared to the published Hubble Space Telescope Wide-Field Camera 3 and Very Large Telescope transit observations by Berta et al. and Bean et al. We argue that the required precision for the distinction may be achievable with currently available instrumentation by stacking 10-15 repeated transit observations. We provide a scaling law that scales our quantitative results to other transiting super-Earths and Neptunes such as HD 97658b, 55 Cnc e, GJ 3470b and GJ 436b. The analysis in this work is performed using an improved version of our Bayesian atmospheric retrieval framework. The new framework not only constrains the gas composition and cloud/haze parameters, but also determines our confidence in having detected molecules and cloud/haze species through Bayesian model comparison. Using the Bayesian tool, we demonstrate quantitatively that the subtle transit depth variation in the Berta et al. data is not sufficient to claim the detection of water absorption.

79 ASTRONOMY AND ASTROPHYSICS↗

Efficient high-dimensional variational data assimilation with machine-learned reduced-order models

Abstract. Data assimilation (DA) in geophysical sciences remains the cornerstone of robust forecasts from numerical models. Indeed, DA plays a crucial role in the quality of numerical weather prediction and is a crucial building block that has allowed dramatic improvements in weather forecasting over the past few decades. DA is commonly framed in a variational setting, where one solves an optimization problem within a Bayesian formulation using raw model forecasts as a prior and observations as likelihood. This leads to a DA objective function that needs to be minimized, where the decision variables are the initial conditions specified to the model. In traditional DA, the forward model is numerically and computationally expensive. Here we replace the forward model with a low-dimensional, data-driven, and differentiable emulator. Consequently, gradients of our DA objective function with respect to the decision variables are obtained rapidly via automatic differentiation. We demonstrate our approach by performing an emulator-assisted DA forecast of geopotential height. Our results indicate that emulator-assisted DA is faster than traditional equation-based DA forecasts by 4 orders of magnitude, allowing computations to be performed on a workstation rather than a dedicated high-performance computer. In addition, we describe accuracy benefits of emulator-assisted DA when compared to simply using the emulator for forecasting (i.e., without DA). Our overall formulation is denoted AIEADA (Artificial Intelligence Emulator-Assisted Data Assimilation).

58 GEOSCIENCES↗

Unbinned extraction of $γ$ from $B\to DK$ with normalizing flows

We introduce an unbinned method for extracting the CKM angle $γ$ from the decay chain $B^\pm \to (D \to K_S π^+ π^-) K^\pm$ using normalizing flows (NFs). The NFs, trained on $D$ decay data, learn a faithful continuous representation of the amplitude and strong phase variation over the $D\to K_Sπ^+π^-$ Dalitz plot whose fidelity improves with increased data sample sizes. With this input, the $B$ decay data can be used to extract the parameters $r_B$, $δ_B$, and $γ$. We test the method on Monte Carlo generated data, where it successfully recovers the injected value of $γ$ within uncertainties. The present implementation propagates statistical uncertainties from finite training data via an ensemble of independently trained flows, and does not attempt to capture the effects of systematic experimental errors. We explore two versions of the method that differ in how the trigonometric constraint on phase variation is encoded, and comment on the possible extension to Bayesian NFs, which would provide direct uncertainty estimates on the learned densities without requiring ensemble training.

Grossman, Yuval [Cornell U., LEPP]↗

A Variational Inference Approach to Inverse Problems with Gamma Hyperpriors

Hierarchical models with gamma hyperpriors provide a flexible, sparse-promoting framework to bridge L 1 and L 2 regularizations in Bayesian formulations to inverse problems. Despite the Bayesian motivation for these models, existing methodologies are limited to maximum a posteriori estimation. The potential to perform uncertainty quantification has not yet been realized. This paper introduces a variational iterative alternating scheme for hierarchical inverse problems with gamma hyperpriors. The proposed variational inference approach yields accurate reconstruction, provides meaningful uncertainty quantification, and is easy to implement. In addition, it lends itself naturally to conduct model selection for the choice of hyperparameters. Here, we illustrate the performance of our methodology in several computed examples, including a deconvolution problem and sparse identification of dynamical systems from time series data.

Bayesian shrinkage↗

B-PINNs: Bayesian Physics-informal Neural Networks for Forward and Inverse PDE Problems with Noisy Data

We propose a Bayesian physics-informed neural network (B-PINN) to solve both forward and inverse nonlinear problems described by partial differential equations (PDEs) and noisy data. In this Bayesian framework, the Bayesian neural network (BNN) combined with a PINN for PDEs serves as the prior while the Hamiltonian Monte Carlo (HMC) or the variational inference (VI) could serve as an estimator of the posterior. B-PINNs make use of both physical laws and scattered noisy measurements to provide predictions and quantify the aleatoric uncertainty arising from the noisy data in the Bayesian framework. Compared with PINNs, in addition to uncertainty quantification, B-PINNs obtain more accurate predictions in scenarios with large noise due to their capability of avoiding overfitting. We conduct a systematic comparison between the two different approaches for the B-PINNs posterior estimation (i.e., HMC or VI), along with dropout used for quantifying uncertainty in deep neural networks. Our experiments show that HMC is more suitable than VI with mean field Gaussian approximation for the B-PINNs posterior estimation, while dropout employed in PINNs can hardly provide accurate predictions with reasonable uncertainty. Finally, we replace the BNN in the prior with a truncated Karhunen-Loève (KL) expansion combined with HMC or a deep normalizing flow (DNF) model as posterior estimators. The KL is as accurate as BNN and much faster but this framework cannot be easily extended to high-dimensional problems unlike the BNN based framework.

Non linear PDEs, Noisy data, Bysian physics inform↗

TYPE Ia SUPERNOVA LIGHT-CURVE INFERENCE: HIERARCHICAL BAYESIAN ANALYSIS IN THE NEAR-INFRARED

We present a comprehensive statistical analysis of the properties of Type Ia supernova (SN Ia) light curves in the near-infrared using recent data from Peters Automated InfraRed Imaging TELescope and the literature. We construct a hierarchical Bayesian framework, incorporating several uncertainties including photometric error, peculiar velocities, dust extinction, and intrinsic variations, for principled and coherent statistical inference. SN Ia light-curve inferences are drawn from the global posterior probability of parameters describing both individual supernovae and the population conditioned on the entire SN Ia NIR data set. The logical structure of the hierarchical model is represented by a directed acyclic graph. Fully Bayesian analysis of the model and data is enabled by an efficient Markov Chain Monte Carlo algorithm exploiting the conditional probabilistic structure using Gibbs sampling. We apply this framework to the JHK{sub s} SN Ia light-curve data. A new light-curve model captures the observed J-band light-curve shape variations. The marginal intrinsic variances in peak absolute magnitudes are sigma(M{sub J}) = 0.17 +- 0.03, sigma(M{sub H}) = 0.11 +- 0.03, and sigma(M{sub Ks}) = 0.19 +- 0.04. We describe the first quantitative evidence for correlations between the NIR absolute magnitudes and J-band light-curve shapes, and demonstrate their utility for distance estimation. The average residual in the Hubble diagram for the training set SNe at cz > 2000kms{sup -1} is 0.10 mag. The new application of bootstrap cross-validation to SN Ia light-curve inference tests the sensitivity of the statistical model fit to the finite sample and estimates the prediction error at 0.15 mag. These results demonstrate that SN Ia NIR light curves are as effective as corrected optical light curves, and, because they are less vulnerable to dust absorption, they have great potential as precise and accurate cosmological distance indicators.

79 ASTRONOMY AND ASTROPHYSICS↗

Soil moisture estimation using tomographic ground penetrating radar in a MCMC–Bayesian framework

In this paper, we focus on a hydrogeological inverse problem specifically targeting monitoring soil moisture variations using tomographic ground penetrating radar (GPR) travel time data. Technical challenges exist in the inversion of GPR tomographic data for handling non-uniqueness, nonlinearity and high-dimensionality of unknowns. We have developed a new method for estimating soil moisture fields from crosshole GPR data. It uses a pilot-point method to provide a low-dimensional representation of the relative dielectric permittivity field of the soil, which is the primary object of inference: the field can be converted to soil moisture using a petrophysical model. We integrate a multi-chain Markov chain Monte Carlo (MCMC)–Bayesian inversion framework with the pilot point concept, a curved-ray GPR travel time model, and a sequential Gaussian simulation algorithm, for estimating the dielectric permittivity at pilot point locations distributed within the tomogram, as well as the corresponding geostatistical parameters (i.e., spatial correlation range). We infer the dielectric permittivity as a probability density function, thus capturing the uncertainty in the inference. The multi-chain MCMC enables addressing high-dimensional inverse problems as required in the inversion setup. The method is scalable in terms of number of chains and processors, and is useful for computationally demanding Bayesian model calibration in scientific and engineering problems. The proposed inversion approach can successfully approximate the posterior density distributions of the pilot points, and capture the true values. The computational efficiency, accuracy, and convergence behaviors of the inversion approach were also systematically evaluated, by comparing the inversion results obtained with different levels of noises in the observations, increased observational data, as well as increased number of pilot points.

54 ENVIRONMENTAL SCIENCES↗

Defining Change Thresholds: What Change Is Outside Typical Sources of Variation?

Researchers often have a difficult time defining meaningful thresholds for change. We sometimes identify subtle changes but what amount of change is beyond typical sources of variation? This is especially complicated when trying to understand new disease pathogenesis like the constellation of eye changes leading to Spaceflight-associated Neuro-ocular Syndrome (SANS). To support decision makers in defining minimal meaningful change, we used a Bayesian hierarchical model to estimate innate sources of variability such as natural day to day variation. Healthy subjects were recruited and imaged with MRI, OCT, and US on separate days and measured by several technicians. Models were developed specifying random effects for the sources of variation – between left and right eyes, within-individuals over time, between raters, and finally between individuals. This allowed us to find the posterior distribution for the total typical variation, within an eye, which we use to define a threshold where change beyond typical sources of variation is likely. This threshold is now used as our earliest indicator of systematic increase in Total Retinal Thickness (a precursor to optic disc edema).

Millennia Young↗