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 73 records · Page 4

Variational inference of effective range parameters for 3 He− 4 He scattering

We use two different methods, Monte Carlo sampling and variational inference (VI), to perform a Bayesian calibration of the effective-range parameters in 3 He– 4 He elastic scattering. The parameters are calibrated to data from a recent set of 3 He– 4 He elastic scattering differential cross section measurements. Analysis of these data for E lab ≤ 4.3 MeV yields a unimodal posterior for which both methods obtain the same structure. However, the effective-range expansion amplitude does not account for the 7/2 − state of 7 Be so, even after calibration, the description of data at the upper end of this energy range is poor. The data up to E lab = 2.6 MeV can be well described, but calibration to this lower-energy subset of the data yields a bimodal posterior. After adapting VI to treat such a multi-modal posterior we find good agreement between the VI results and those obtained with parallel-tempered Monte Carlo sampling.

effective field theory↗

Fracture detection and imaging through relative seismic velocity changes using distributed acoustic sensing and ambient seismic noise

Fracture systems are important pathways for fluid and solute transport and exert a critical influence on the hydraulic properties of aquifers and reservoirs. Therefore, detailed knowledge of fracture locations, connections, and evolution is crucial for both groundwater and energy applications (e.g., enhanced geothermal, oil and gas recovery, carbon sequestration, and wastewater injection). The innovative combination of distributed acoustic sensing (DAS) and ambient seismic noise techniques has the potential to detect and characterize fracture systems at high-spatial and temporal resolution without an active source. To test this, we conducted a multiphysics field experiment at Blue Canyon Dome, New Mexico. A novel energetic material developed by Sandia National Laboratories was used to generate fractures in two separate stimulations. Ambient noise was recorded before and after each stimulation using fiber-optic cables installed in the outer annulus of four boreholes surrounding the stimulation hole at a radius of 1.2 m. The Python package MSNoise was used to compute crosscorrelations and measure changes in velocity between each time period relative to the initial (prestimulation) time period. The majority of channel pairs showed a velocity reduction (average –3% relative velocity change) following both stimulations. We used a 3D Bayesian tomography approach to resolve spatial variations by utilizing differences between channel pairs. Results showed that the greatest velocity reduction was concentrated near the center of the test area and suggested the presence of a near-vertical fracture, oriented northeast to southwest for depths >19 m below ground surface and extending slightly to the southwest corner. Finally, these results were generally consistent with crosshole seismic tomography time-lapse images. DAS technology provides valuable sensing capability and — when used with a passive seismic approach — shows great promise for monitoring and characterization of fractured-rock systems.

58 GEOSCIENCES↗

Multifidelity Ensemble Kalman Filtering Using Surrogate Models Defined by Theory-Guided Autoencoders

Data assimilation is a Bayesian inference process that obtains an enhanced understanding of a physical system of interest by fusing information from an inexact physics-based model, and from noisy sparse observations of reality. The multifidelity ensemble Kalman filter (MFEnKF) recently developed by the authors combines a full-order physical model and a hierarchy of reduced order surrogate models in order to increase the computational efficiency of data assimilation. The standard MFEnKF uses linear couplings between models, and is statistically optimal in case of Gaussian probability densities. This work extends the MFEnKF into to make use of a broader class of surrogate model such as those based on machine learning methods such as autoencoders non-linear couplings in between the model hierarchies. We identify the right-invertibility property for autoencoders as being a key predictor of success in the forecasting power of autoencoder-based reduced order models. We propose a methodology that allows us to construct reduced order surrogate models that are more accurate than the ones obtained via conventional linear methods. Numerical experiments with the canonical Lorenz'96 model illustrate that nonlinear surrogates perform better than linear projection-based ones in the context of multifidelity ensemble Kalman filtering. We additionality show a large-scale proof-of-concept result with the quasi-geostrophic equations, showing the competitiveness of the method with a traditional reduced order model-based MFEnKF.

97 MATHEMATICS AND COMPUTING↗

A support vector machine classifier reduces interscanner variation in the HRCT classification of regional disease pattern in diffuse lung disease: Comparison to a Bayesian classifier

Purpose: To investigate the effect of using different computed tomography (CT) scanners on the accuracy of high-resolution CT (HRCT) images in classifying regional disease patterns in patients with diffuse lung disease, support vector machine (SVM) and Bayesian classifiers were applied to multicenter data. Methods: Two experienced radiologists marked sets of 600 rectangular 20 Multiplication-Sign 20 pixel regions of interest (ROIs) on HRCT images obtained from two scanners (GE and Siemens), including 100 ROIs for each of local patterns of lungs-normal lung and five of regional pulmonary disease patterns (ground-glass opacity, reticular opacity, honeycombing, emphysema, and consolidation). Each ROI was assessed using 22 quantitative features belonging to one of the following descriptors: histogram, gradient, run-length, gray level co-occurrence matrix, low-attenuation area cluster, and top-hat transform. For automatic classification, a Bayesian classifier and a SVM classifier were compared under three different conditions. First, classification accuracies were estimated using data from each scanner. Next, data from the GE and Siemens scanners were used for training and testing, respectively, and vice versa. Finally, all ROI data were integrated regardless of the scanner type and were then trained and tested together. All experiments were performed based on forward feature selection and fivefold cross-validation with 20 repetitions. Results: For each scanner, better classification accuracies were achieved with the SVM classifier than the Bayesian classifier (92% and 82%, respectively, for the GE scanner; and 92% and 86%, respectively, for the Siemens scanner). The classification accuracies were 82%/72% for training with GE data and testing with Siemens data, and 79%/72% for the reverse. The use of training and test data obtained from the HRCT images of different scanners lowered the classification accuracy compared to the use of HRCT images from the same scanner. For integrated ROI data obtained from both scanners, the classification accuracies with the SVM and Bayesian classifiers were 92% and 77%, respectively. The selected features resulting from the classification process differed by scanner, with more features included for the classification of the integrated HRCT data than for the classification of the HRCT data from each scanner. For the integrated data, consisting of HRCT images of both scanners, the classification accuracy based on the SVM was statistically similar to the accuracy of the data obtained from each scanner. However, the classification accuracy of the integrated data using the Bayesian classifier was significantly lower than the classification accuracy of the ROI data of each scanner. Conclusions: The use of an integrated dataset along with a SVM classifier rather than a Bayesian classifier has benefits in terms of the classification accuracy of HRCT images acquired with more than one scanner. This finding is of relevance in studies involving large number of images, as is the case in a multicenter trial with different scanners.

60 APPLIED LIFE SCIENCES↗

Brittleness and Bayesian Inference

In a recent report, “Bayesian Brittleness: Why no Bayesian model is ‘good enough,’” (arXiv, 1304.6772v1) the authors prove rigorously that under a seemingly natural class of priors Π, the posterior expectation of a function Φ(θ) can achieve essentially any value that Φ(θ) 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 problem is that the priors leading to the extreme results depend on the observed data, which means they are not valid priors. The corresponding posteriors are also invalid, so the extreme variation in these posteriors does not imply that Bayesian inference is brittle. The report, however, provides a new framework which may be used for any assumed class of priors Π, and which we explain in detail. This framework may be a useful new tool for deriving quantitative robustness results.

97 MATHEMATICS AND COMPUTING↗

Ensemble variational Fokker-Planck methods for data assimilation

Particle flow filters solve Bayesian inference problems by smoothly transforming a set of particles into samples from the posterior distribution. Particles move in state space under the flow of an McKean-Vlasov-Itˆo process. This work introduces the Variational Fokker-Planck (VFP) framework for data assimilation, a general approach that includes previously known particle flow filters as special cases. The McKean-Vlasov-Itˆo process that transforms particles is defined via an optimal drift that depends on the selected diffusion term. It is established that the underlying probability density - sampled by the ensemble of particles - converges to the Bayesian posterior probability density. For a finite number of particles the optimal drift contains a regularization term that nudges particles toward becoming independent random variables. Based on this analysis, we derive computationally-feasible approximate regularization approaches that penalize the mutual information between pairs of particles, and avoid particle collapse. Moreover, the diffusion plays a role akin to a particle rejuvenation approach that aims to alleviate particle collapse. The VFP framework is very flexible. Different assumptions on prior and intermediate probability distributions can be used to implement the optimal drift, and localization and covariance shrinkage can be applied to alleviate the curse of dimensionality. A robust implicit-explicit method is discussed for the efficient integration of stiff McKean- Vlasov-Itˆo processes. Here, the effectiveness of the VFP framework is demonstrated on three progressively more challenging test problems, namely the Lorenz ’63, Lorenz ’96 and the quasi-geostrophic equations.

97 MATHEMATICS AND COMPUTING↗

A differentiable simulation package for performing inference of synchrotron-radiation-based diagnostics

The direction of particle accelerator development is ever-increasing beam quality, currents and repetition rates. This poses a challenge to traditional diagnostics that directly intercept the beam due to the mutual destruction of both the beam and the diagnostic. An alternative approach is to infer beam parameters non-invasively from the synchrotron radiation emitted in bending magnets. However, inferring the beam distribution from a measured radiation pattern is a complex and computationally expensive task. To address this challenge we present SYRIPY ( SYnchrotron Radiation In PYthon ), a software package intended as a tool for performing inference of synchrotron-radiation-based diagnostics. SYRIPY has been developed using PyTorch , which makes it both differentiable and able to leverage the high performance of GPUs, two vital characteristics for performing statistical inference. The package consists of three modules: a particle tracker, Lienard–Wiechert solver and Fourier optics propagator, allowing start-to-end simulation of synchrotron radiation detection to be carried out. SYRIPY has been benchmarked against SRW , the prevalent numerical package in the field, showing good agreement and up to a 50× speed improvement. Finally, we have demonstrated how SYRIPY can be used to perform Bayesian inference of beam parameters using stochastic variational inference.

43 PARTICLE ACCELERATORS↗

Uncertainties in Parameters Estimated with Neural Networks: Application to Strong Gravitational Lensing

In Hezaveh et al. (2017) we showed that deep learning can be used for model parameter estimation and trained convolutional neural networks to determine the parameters of strong gravitational lensing systems. Here we demonstrate a method for obtaining the uncertainties of these parameters. We review the framework of variational inference to obtain approximate posteriors of Bayesian neural networks and apply it to a network trained to estimate the parameters of the Singular Isothermal Ellipsoid plus external shear and total flux magnification. We show that the method can capture the uncertainties due to different levels of noise in the input data, as well as training and architecture-related errors made by the network. To evaluate the accuracy of the resulting uncertainties, we calculate the coverage probabilities of marginalized distributions for each lensing parameter. By tuning a single hyperparameter, the dropout rate, we obtain coverage probabilities approximately equal to the confidence levels for which they were calculated, resulting in accurate and precise uncertainty estimates. Our results suggest that neural networks can be a fast alternative to Monte Carlo Markov Chains for parameter uncertainty estimation in many practical applications, allowing more than seven orders of magnitude improvement in speed.

Perreault Levasseur, Laurence↗

Bayesian modeling of inconsistent plastic response due to material variability

The advent of fabrication techniques such as additive manufacturing has focused attention on the considerable variability of material response due to defects and other microstructural aspects. This variability motivates the development of an enhanced design methodology that incorporates inherent material variability to provide robust predictions of performance. In this work, we develop plasticity models capable of representing the distribution of mechanical responses observed in experiments using traditional plasticity models of the mean response and recently developed uncertainty quantification (UQ) techniques. To account for material response variability through variations in physical parameters, we adapt a recent Bayesian embedded modeling error calibration technique. We use Bayesian model selection to determine the most plausible of a variety of plasticity models and the optimal embedding of parameter variability. To expedite model selection, we develop an adaptive importance-sampling-based numerical integration scheme to compute the Bayesian model evidence. In conclusion, we demonstrate that the new framework provides predictive realizations that are superior to more traditional ones, and how these UQ techniques can be used in model selection and assessing the quality of calibrated physical parameters.

36 MATERIALS SCIENCE↗

A rotorcraft in-flight ice detection framework using computational aeroacoustics and Bayesian neural networks

Abstract This work develops a novel ice detection framework specifically suitable for rotorcraft using computational aeroacoustics and Bayesian neural networks. In an offline phase of the work, the acoustic signature of glaze and rime ice shapes on an oscillating wing are computed. In addition, the aerodynamic performance indicators corresponding to the ice shapes are also monitored. These performance indicators include the lift, drag, and moment coefficients. A Bayesian neural network is subsequently trained using projected Stein variational gradient descent to create a mapping from the acoustic signature generated by the iced wings to predict their performance indicators along with quantified uncertainty that is highly important for time- and safety-critical decision-making scenarios. While the training is carried out fully offline, usage of the Bayesian neural network to make predictions can be conducted rapidly online allowing for an ice detection system that can be used in real time and in-flight.

42 ENGINEERING↗

CONCURRENT, CONDENSED STEIN VARIATIONAL GRADIENT DESCENT FOR UNCERTAINTY QUANTIFICATION OF NEURAL NETWORKS

In this work, we propose a Stein variational gradient descent (SVGD) method to concurrently sparsify, train, and provide uncertainty quantification (UQ) of a complexly parameterized model, such as a neural network (NN). It employs a graph reconciliation and condensation process to reduce complexity and increase similarity in the Stein ensemble of parameterizations. Therefore, the proposed concurrent, condensed SVGD (ccSVGD) method can provide UQ on parameters, not just outputs. Furthermore, the parameter reduction speeds up the convergence of the Stein gradient descent as it reduces the combinatorial complexity by aligning and differentiating the sensitivity to parameters. These properties are demonstrated with an illustrative example and an application to a mechanical response representation problem in solid mechanics.

42 ENGINEERING↗

Application of a Geochemically Informed Leak Detection (GILD) Model to CO 2 Injection Sites on the United States Gulf Coast

The Gulf Coast region possesses great potential for CO 2 enhanced oil recovery (EOR) and CO 2 storage. A geochemically informed leak detection (GILD) model has been applied to CO 2 injection sites on the Gulf Coast with considerations of measurement variability. The Jasper aquifer in Montgomery County, Texas, was chosen to demonstrate the method. Based on background data from wells in the area, combinations of mineral and fluid compositions were used to create 23 scenarios for the geochemical model. The output from the geochemical model was used to identify sensitive monitoring species, and response functions were generated for these as a function of the CO 2 leakage concentration. The sources of measurement variability for background conditions were characterized from the Jasper aquifer background data, and then normalized using the coefficient of variation of each species across the monitoring wells. Bayesian belief network (BBN) models were constructed, and measurement variability of different levels were added to compare leak detection probabilities. Increasing measurement variability decreased the power to detect a leak of a given size. For a moderately high CO 2 concentration of 0.2 mol/kg, the probability of detecting this leakage effect using pH as the monitoring variable in an aquifer with calcite decreases from 98% (no measurement variability) to 61% (medium variability) to 33% (high variability). The loss in power of the sampling protocol with increasing measurement variability is similar in magnitude when Ca 2+ or HCO 3 - is used as the monitoring parameter, but only for aquifers with calcite.

58 GEOSCIENCES↗

Hydro-scaling of direct-drive cylindrical implosions at the OMEGA and the National Ignition Facility

Deceleration-phase Rayleigh–Taylor instability (RTI) growth during inertial confinement fusion capsule implosions significantly affects the performance as it mixes cold ablator material into the fuel. Precise measurements of such instability growth are essential for both validating the existing simulation codes and improving our predictive capability. RTI measurements on the inner surface of a spherical shell are limited and are often inferred indirectly at limited convergence. In contrast, cylindrical implosions allow for direct diagnostic access to the converging interface by imaging down the cylinder axis while retaining the effects of convergence. We have performed direct-drive cylindrical implosion experiments at both the OMEGA and the NIF laser facilities using scaled targets. RTI growth is demonstrated to be scale-invariant between the cylindrical targets at OMEGA and similar targets at the NIF, which are scaled up by a factor of three in the radial dimension. Single-mode (m = 0) instability growth factors of ~14 are measured at a convergence ratio (CR) ~ 2.5 with nearly identical mode growth at both scales. The measurements are in agreement with xRAGE radiation-hydrodynamics simulations. In addition, we have developed the Bayesian-inference-engine method to account for the variations in the target alignment, magnification, and the parallax effect in the measurement, allowing a more precise comparison between the experimental data and the simulations.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Quantification of neural networks uncertainties with applications to SAFARI-1 axial neutron flux profiles

Deep Neural Networks (DNNs) have been widely used as a data-driven modelling tool in nuclear engineering. However, as a Machine Learning model, Artificial Neural Network (ANN) predictions are subjected to uncertainties originating from the noise in training data, incomplete coverage of the domain, and imperfect neural network architectures. In this work, we target at quantifying the prediction/approximation uncertainties of ANNs using Monte Carlo Dropout (MCD), as well as Bayesian Neural Networks (BNNs) which are solved by variational inference. With a demonstration problem in which neural networks are used to predict the assembly axial neutron flux profiles, the results have shown that the three different neural network models (regular DNNs, DNNs solved with MCD and BNNs) can produce results that agree very well among each other and with the measurement data, on cycles that are not used in the training process. Besides the excellent generalization capability, the uncertainty bands produced by MCD and BNN agree very well, and in general, they can fully envelop the noisy measurement data points. (authors)

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Observation of CH4 and other Non-CO2 Green House Gas Emissions from California

In 2006, California passed the landmark assembly bill AB-32 to reduce California's emissions of greenhouse gases (GHGs) that contribute to global climate change. AB-32 commits California to reduce total GHG emissions to 1990 levels by 2020, a reduction of 25 percent from current levels. To verify that GHG emission reductions are actually taking place, it will be necessary to measure emissions. We describe atmospheric inverse model estimates of GHG emissions obtained from the California Greenhouse Gas Emissions Measurement (CALGEM) project. In collaboration with NOAA, we are measuring the dominant long-lived GHGs at two tall-towers in central California. Here, we present estimates of CH{sub 4} emissions obtained by statistical comparison of measured and predicted atmospheric mixing ratios. The predicted mixing ratios are calculated using spatially resolved a priori CH{sub 4} emissions and surface footprints, that provide a proportional relationship between the surface emissions and the mixing ratio signal at tower locations. The footprints are computed using the Weather Research and Forecast (WRF) coupled to the Stochastic Time-Inverted Lagrangian Transport (STILT) model. Integral to the inverse estimates, we perform a quantitative analysis of errors in atmospheric transport and other factors to provide quantitative uncertainties in estimated emissions. Regressions of modeled and measured mixing ratios suggest that total CH{sub 4} emissions are within 25% of the inventory estimates. A Bayesian source sector analysis obtains posterior scaling factors for CH{sub 4} emissions, indicating that emissions from several of the sources (e.g., landfills, natural gas use, petroleum production, crops, and wetlands) are roughly consistent with inventory estimates, but livestock emissions are significantly higher than the inventory. A Bayesian 'region' analysis is used to identify spatial variations in CH{sub 4} emissions from 13 sub-regions within California. Although, only regions near the tower are significantly constrained by the tower measurements, CH{sub 4} emissions from the south Central Valley appear to be underestimated in a manner consistent with the under-prediction of livestock emissions. Finally, we describe a pseudo-experiment using predicted CH{sub 4} signals to explore the uncertainty reductions that might be obtained if additional measurements were made by a future network of tall-tower stations spread over California. These results show that it should be possible to provide high-accuracy estimates of surface CH{sub 4} emissions for multiple regions as a means to verify future emissions reductions.

54 ENVIRONMENTAL SCIENCES↗

Learning from crowds with variational Gaussian processes

Solving a supervised learning problem requires to label a training set. This task is traditionally performed by an expert, who provides a label for each sample. The proliferation of social web services (e.g., Amazon Mechanical Turk) has introduced an alternative crowdsourcing approach. Anybody with a computer can register in one of these services and label, either partially or completely, a dataset. The effort of labeling is then shared between a great number of annotators. However, this approach introduces scientifically challenging problems such as combining the unknown expertise of the annotators, handling disagreements on the annotated samples, or detecting the existence of spammer and adversarial annotators. All these problems require probabilistic sound solutions which go beyond the naive use of majority voting plus classical classification methods. In this work we introduce a new crowdsourcing model and inference procedure which trains a Gaussian Process classifier using the noisy labels provided by the annotators. Variational Bayes inference is used to estimate all unknowns. The proposed model can predict the class of new samples and assess the expertise of the involved annotators. Moreover, the Bayesian treatment allows for a solid uncertainty quantification. Since when predicting the class of a new sample we might have access to some annotations for it, we also show how our method can naturally incorporate this additional information. Furthermore, a comprehensive experimental section evaluates the proposed method with synthetic and real experiments, showing that it consistently outperforms other state-of-the-art crowdsourcing approaches.

97 MATHEMATICS AND COMPUTING↗

Designing an Optimal Sensor Network via Minimizing Information Loss

Optimal experimental design is a classic topic in statistics, with many well-studied problems, applications, and solutions. The design problem we study is the placement of sensors to monitor spatiotemporal processes, explicitly accounting for the temporal dimension in our modeling and optimization. We observe that recent advancements in computational sciences often yield large datasets based on physics-based simulations, which are rarely leveraged in experimental design. We introduce a novel model-based sensor placement criterion, along with a highly-efficient optimization algorithm, which integrates physics-based simulations and Bayesian experimental design principles to identify sensor networks that “minimize information loss” from simulated data. Our technique relies on sparse variational inference and (separable) Gauss-Markov priors, and thus may adapt many techniques from Bayesian experimental design. We validate our method through a case study monitoring air temperature in Phoenix, Arizona, using state-of-the-art physics-based simulations. Our results show our framework to be superior to random or quasi-random sampling, particularly with a limited number of sensors. We conclude by discussing practical considerations and implications of our framework, including more complex modeling tools and real-world deployments.

54 ENVIRONMENTAL SCIENCES↗

Optimizing Optical Searches for Supermassive Black Hole Binaries in Active Galactic Nuclei Light Curves: Fourier versus Bayesian Periodicity Detection

Simulations predict that supermassive black hole binaries (SMBHBs) will exhibit periodic brightness variations that may exceed the stochastic variability intrinsic to active galactic nuclei (AGN). In this paper, we simulate SMBHBs with damped random walk (DRW) AGN variability and an added sinusoidal signal from the orbital motion, and test three methods—a generalized Lomb–Scargle periodogram (GLSP), a nested Bayesian sampler (NBS), and a weighted wavelet z-transform (or WWZ)—to determine which is best at recovering the periodicity. Our simulated light curves follow the properties of the Catalina Real-Time Transient Survey (or CRTS), Legacy Survey of Space and Time (LSST), and Zwicky Transient Facility (ZTF) to best inform current and future SMBHB searches. We map a broad range of parameter space and identify which DRW-only light curves best mimic periodicity and pass each method’s model selection. The NBS performs best at detecting periodicity and filtering out DRW-only light curves. Combined candidate selection with both the NBS and GLSP significantly reduces false-positive rates (FPRs) with marginal impact on true-positive rates (TPRs). With this joint model selection pipeline, we find the lowest FPRs in ZTF-like simulations and the highest detection rates in LSST-like simulations. Using a modified computation of the false-alarm probability with GLSP, we efficiently triage LSST AGN light curves (∼10 7 light curves in ∼10–30 hr) and achieve TPRs and FPRs of ∼40% and ∼0.5%, respectively.

Banaszak, Sebastian M. [Vanderbilt Univ., Nashvill↗