Search NASA⌕ Search

SEARCH · Search NASA

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
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

Particle image velocimetry analysis with simultaneous uncertainty quantification using Bayesian neural networks

Particle image velocimetry (PIV) is an effective tool in experimental fluid mechanics for extracting flow fields from images. Recently, convolutional neural networks (CNNs) have been used to perform PIV analysis with accuracy on par with classical methods. Here we extend the use of CNNs to analyze PIV data while providing simultaneous uncertainty quantification on the inferred flow field. The method we apply in this paper is a Bayesian convolutional neural network (BCNN) which learns distributions of the CNN weights through variational Bayes. In order to demonstrate the utility of BCNNs for the PIV task, we compare the performance of three distinct BCNN models with simple architectures. The first network estimates flow velocity from image interrogation regions only. Our second model learns to infer velocity from both the image interrogation regions and interrogation region cross-correlation maps. Finally, our best performing network infers velocities from interrogation region cross-correlation maps only. We find that BCNNs using interrogation region cross-correlation maps as inputs perform better than those using interrogation windows only as inputs and discuss reasons why this may be the case. Additionally, we test the best performing BCNN on a full synthetic test image pair and a real image pair from the 1st International PIV Challenge. We show that ~98% of true particle displacements from the full synthetic image pair can be captured within the BCNN's 95% confidence intervals, and that the BCNN's performance on the real image pair is quantitatively similar to that of algorithms tested in the 1st International PIV Challenge. Finally, we show that BCNNs can be generalized to be used with multi-pass PIV algorithms with a moderate loss in accuracy, which may be overcome by future work on finetuning and training schemes. So to our knowledge, this is the first use of Bayesian neural networks to perform PIV.

47 OTHER INSTRUMENTATION↗
DOE OSTI2025

An investigation on machine learning predictive accuracy improvement and uncertainty reduction using VAE-based data augmentation

The confluence of ultrafast computers with large memory, rapid progress in Machine Learning (ML) algorithms, and the availability of large datasets place multiple engineering fields at the threshold of dramatic progress. However, a unique challenge in nuclear engineering is data scarcity because experimentation on nuclear systems is usually more expensive and time-consuming than most other disciplines. One potential way to resolve the data scarcity issue is deep generative learning, which uses certain ML models to learn the underlying distribution of existing data and generate synthetic samples that resemble the real data. In this way, one can significantly expand the dataset to train more accurate predictive ML models. In this study, our objective is to evaluate the effectiveness of data augmentation using variational autoencoder (VAE)-based deep generative models. We investigated whether the data augmentation leads to improved accuracy in the predictions of a deep neural network (DNN) model trained using the augmented data. Additionally, the DNN prediction uncertainties are quantified using Bayesian Neural Networks (BNN) and conformal prediction (CP) to assess the impact on predictive uncertainty reduction. To test the proposed methodology, we used TRACE simulations of steady-state void fraction data based on the NUPEC Boiling Water Reactor Full-size Fine-mesh Bundle Test (BFBT) benchmark. Here, we found that augmenting the training dataset using VAEs has improved the DNN model’s predictive accuracy, improved the prediction confidence intervals, and reduced the prediction uncertainties.

Bayesian neural network↗
DOE OSTI2016

The Casualidad method for uncertainty evaluation of best-estimate system thermal-hydraulics calculations

The present paper deals with the description of the salient features of three independent approaches for estimating uncertainties associated with predictions of complex system codes. The first approach is the 'standard' one and the most used at the industrial level: it is based upon the selection of input uncertain parameters, on assigning related ranges of variations and, possibly, PDF (Probability Density Functions) and on performing a suitable number of code runs to get the combined effect of variation on the results. In the second approach the uncertainty derives from the comparison between relevant measured data and results of corresponding code calculations. The third approach is based on the Bayesian inference technique and on the availability of experimental data by which computer model predictions can be improved and the ranges of variation of (in theory) 'all' input parameters can be characterized. More details are provided in respect with the third approach that has been named CASUALIDAD (Code with the capability of Adjoint Sensitivity and Uncertainty Analysis by Internal Data Adjustment and assimilation). (authors)

42 ENGINEERING↗
DOE OSTI2021

Bayesian model averaging for analysis of lattice field theory results

Statistical modeling is a key component in the extraction of physical results from lattice field theory calculations. Although the general models used are often strongly motivated by physics, many model variations can frequently be considered for the same lattice data. Model averaging, which amounts to a probability-weighted average over all model variations, can incorporate systematic errors associated with model choice without being overly conservative. We discuss the framework of model averaging from the perspective of Bayesian statistics, and give useful formulae and approximations for the particular case of least-squares fitting, commonly used in modeling lattice results. In addition, we frame the common problem of data subset selection (e.g. choice of minimum and maximum time separation for fitting a two-point correlation function) as a model selection problem and study model averaging as a straightforward alternative to manual selection of fit ranges. Numerical examples involving both mock and real lattice data are given.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2020

A novel CMB component separation method: hierarchical generalized morphological component analysis

ABSTRACT We present a novel technique for cosmic microwave background (CMB) foreground subtraction based on the framework of blind source separation. Inspired by previous work incorporating local variation to generalized morphological component analysis (GMCA), we introduce hierarchical GMCA (HGMCA), a Bayesian hierarchical graphical model for source separation. We test our method on Nside = 256 simulated sky maps that include dust, synchrotron, free–free, and anomalous microwave emission, and show that HGMCA reduces foreground contamination by $25{{\ \rm per\ cent}}$ over GMCA in both the regions included and excluded by the Planck UT78 mask, decreases the error in the measurement of the CMB temperature power spectrum to the 0.02–0.03 per cent level at ℓ > 200 (and $\lt 0.26{{\ \rm per\ cent}}$ for all ℓ), and reduces correlation to all the foregrounds. We find equivalent or improved performance when compared to state-of-the-art internal linear combination type algorithms on these simulations, suggesting that HGMCA may be a competitive alternative to foreground separation techniques previously applied to observed CMB data. Additionally, we show that our performance does not suffer when we perturb model parameters or alter the CMB realization, which suggests that our algorithm generalizes well beyond our simplified simulations. Our results open a new avenue for constructing CMB maps through Bayesian hierarchical analysis.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2016

Joint Data Filtering and Labeling Using Gaussian Processes and Alternating Direction Method of Multipliers

Here, sequence labeling aims at assigning a label to every sample of a signal (or pixel of an image) while considering the sequentiality (or vicinity) of the samples. To perform this task, many works in the literature first filter and then label the data. Unfortunately, the filtering, which is performed independently from the labeling, is far from optimal and frequently makes the latter task harder. In this paper, a novel approach that trains a Gaussian process classifier and estimates the coefficients of an optimal filter jointly is presented. The new approach, based on Bayesian modeling and alternating direction method of multipliers (ADMMs) optimization, performs both tasks simultaneously. All unknowns are treated as stochastic variables, which are estimated using variational inference and filtering and labeling are linked with the use of ADMM. In the experimental section, synthetic and real experiments are presented to compare the proposed method with other existing approaches.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Bayesian analysis for estimating statistical parameter distributions of elasto-viscoplastic material models

High temperature design methods rely on constitutive models for inelastic deformation and failure typically calibrated against the mean of experimental data without considering the associated scatter. Variability may arise from the experimental data acquisition process, from heat-to-heat material property variations, or both and need to be accurately captured to predict parameter bounds leading to efficient component design. Applying the Bayesian Markov Chain Monte Carlo (MCMC) method to produce statistical models capturing the underlying uncertainty in the experimental data is an area of ongoing research interest. This work varies aspects of the Bayesian MCMC method and explores their effect on the posterior parameter distributions for a uniaxial elasto-viscoplastic damage model using synthetically generated reference data. From our analysis with the uniaxial inelastic model we determine that an informed prior distribution including different types of test conditions results in more accurate posterior parameter distributions. The parameter posterior distributions, however, do not improve when increasing the number of similar experimental data. Additionally, changing the amount of scatter in the data affects the quality of the posterior distributions, especially for the less sensitive model parameters. Moreover, we perform a sensitivity study of the model parameters against the likelihood function prior to the Bayesian analysis. The results of the sensitivity analysis help to determine the reliability of the posterior distributions and reduce the dimensionality of the problem by fixing the insensitive parameters. The comprehensive study described in this work demonstrates how to efficiently apply the Bayesian MCMC methodology to capture parameter uncertainties in high temperature inelastic material models. Quantifying these uncertainties in inelastic models will improve high temperature engineering design practices and lead to safer, more effective component designs.

42 ENGINEERING↗
DOE OSTI2024

Distributionally Robust Variational Quantum Algorithms With Shifted Noise

Given their potential to demonstrate near-term quantum advantage, variational quantum algorithms (VQAs) have been extensively studied. Although numerous techniques have been developed for VQA parameter optimization, it remains a significant challenge. A practical issue is the high sensitivity of quantum noise to environmental changes, and its propensity to shift in real time. This presents a critical problem as an optimized VQA ansatz may not perform effectively under a different noise environment. For the first time, we explore how to optimize VQA parameters to be robust against unknown shifted noise. We model the noise level as a random variable with an unknown probability density function (PDF), and we assume that the PDF may shift within an uncertainty set. This assumption guides us to formulate a distributionally robust optimization problem, with the goal of finding parameters that maintain effectiveness under shifted noise. We utilize a distributionally robust Bayesian optimization solver for our proposed formulation. This provides numerical evidence in both the Quantum Approximate Optimization Algorithm (QAOA) and the Variational Quantum Eigensolver (VQE) with hardware-efficient ansatz, indicating that we can identify parameters that perform more robustly under shifted noise. We regard this work as the first step towards improving the reliability of VQAs influenced by real-time noise.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2026

Multi-Modal Bayesian Neural Network Surrogates with Conjugate Last-Layer Estimation

As data collection and simulation capabilities advance, multi-modal learning, the task of learning from multiple modalities and sources of data, is becoming an increasingly important area of research. Surrogate models that learn from data of multiple auxiliary modalities to support the modeling of a highly expensive quantity of interest have the potential to aid outer loop applications such as optimization, inverse problems, or sensitivity analyses when multi-modal data are available. We develop two multi-modal Bayesian neural network surrogate models and leverage conditionally conjugate distributions in the last layer to estimate model parameters using stochastic variational inference (SVI). We provide a method to perform this conjugate SVI estimation in the presence of partially missing observations. Here, we demonstrate improved prediction accuracy and uncertainty quantification compared to unimodal surrogate models for both scalar and time series data.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Bayesian Spatially Varying Multi-Regularization Image Deblurring

Many scientific experiments such as those found in astronomy, geology, microbiology, and X-ray radiography require the use of high-energy instruments to capture images. Since blur and noise are inevitably present in any imaging system, the images must be \deblurred" to extract the full information content. Mathematically, image deblurring is an ill-posed inverse problem that requires regularization. The regularization, in turn, has a large effect on the deblurred image: different regularization strengths, and types, lead to drastically different reconstructions. Moreover, many images contain a mixture of smooth and sharp features which suggests the use of multi-regularization, i.e., varying the type of regularization (e.g. Tikhonov or total variation) across the image. We address these issues by formulating the image deblurring problem within a hierarchical Bayesian framework in which we spatially adapt the strength of the regularization and also vary the regularization type across the image. In this way, the image itself, along with corresponding regularization strength at each pixel, are described jointly by a posterior distribution which we can sample by Markov chain Monte Carlo (MCMC) methods. We illustrate our techniques on simplified test problems and apply them to high-energy X-ray images taken at the Nevada National Security Site. Numerical tests show that our new method is robustly applicable and increases the quality of the image reconstruction when compared to other (Bayesian) methods.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Assessing variation in faecal glucocorticoid concentrations in gray whales exposed to anthropogenic stressors

Understanding how individual animals respond to stressors behaviourally and physiologically is a critical step towards quantifying long-term population consequences and informing management efforts. Glucocorticoid (GC) metabolite accumulation in various matrices provides an integrated measure of adrenal activation in baleen whales and could thus be used to investigate physiological changes following exposure to stressors. In this study, we measured GC concentrations in faecal samples of Pacific Coast Feeding Group (PCFG) gray whales (Eschrichtius robustus) collected over seven consecutive years to assess the association between GC content and metrics of exposure to sound levels and vessel traffic at different temporal scales, while controlling for contextual variables such as sex, reproductive status, age, body condition, year, time of year and location. We develop a Bayesian Generalized Additive Modelling approach that accommodates the many complexities of these data, including non-linear variation in hormone concentrations, missing covariate values, repeated samples, sampling variability and some hormone concentrations below the limit of detection. Estimated relationships showed large variability, but emerging patterns indicate a strong context-dependency of physiological variation, depending on sex, body condition and proximity to a port. Our results highlight the need to control for baseline hormone variation related to context, which otherwise can obscure the functional relationship between faecal GCs and stressor exposure. Therefore, extensive data collection to determine sources of baseline variation in well-studied populations, such as PCFG gray whales, could shed light on cetacean stress physiology and be used to extend applicability to less-well-studied taxa. GC analyses may offer greatest utility when employed as part of a suite of markers that, in aggregate, provide a multivariate measure of physiological status, better informing estimates of individuals’ health and ultimately the consequences of anthropogenic stressors on populations.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2016

Quantifying the influences of spectral resolution on uncertainty in leaf trait estimates through a Bayesian approach to RTM inversion

The remote monitoring of plant canopies is critically needed for understanding of terrestrial ecosystem mechanics and biodiversity as well as capturing the short- to long-term responses of vegetation to disturbance and climate change. A variety of orbital, sub-orbital, and field instruments have been used to retrieve optical spectral signals and to study different vegetation properties such as plant biochemistry, nutrient cycling, physiology, water status, and stress. Radiative transfer models (RTMs) provide a mechanistic link between vegetation properties and observed spectral features, and RTM spectral inversion is a useful framework for estimating these properties from spectral data. However, existing approaches to RTM spectral inversion are typically limited by the inability to characterize uncertainty in parameter estimates. Here, we introduce a Bayesian algorithm for the spectral inversion of the PROSPECT 5 leaf RTM that is distinct from past approaches in two important ways: First, the algorithm only uses reflectance and does not require transmittance observations, which have been plagued by a variety of measurement and equipment challenges. Second, the output is not a point estimate for each parameter but rather the joint probability distribution that includes estimates of parameter uncertainties and covariance structure. We validated our inversion approach using a database of leaf spectra together with measurements of equivalent water thickness (EWT) and leaf dry mass per unit area (LMA). The parameters estimated by our inversion were able to accurately reproduce the observed reflectance (RMSE VIS = 0.0063, RMSE NIR-SWIR = 0.0098) and transmittance (RMSE VIS = 0.0404, RMSE NIR-SWIR = 0.0551) for both broadleaved and conifer species. Inversion estimates of EWT and LMA for broadleaved species agreed well with direct measurements (CV EWT = 18.8%, CV LMA = 24.5%), while estimates for conifer species were less accurate (CV EWT = 53.2%, CV LMA = 63.3%). To examine the influence of spectral resolution on parameter uncertainty, we simulated leaf reflectance as observed by ten common remote sensing platforms with varying spectral configurations and performed a Bayesian inversion on the resulting spectra. We found that full-range hyperspectral platforms were able to retrieve all parameters accurately and precisely, while the parameter estimates of multispectral platforms were much less precise and prone to bias at high and low values. We also observed that variations in the width and location of spectral bands influenced the shape of the covariance structure of parameter estimates. Lastly, our Bayesian spectral inversion provides a powerful and versatile framework for future RTM development and single- and multi-instrumental remote sensing of vegetation.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Prime VI

SAND2025-03757O Prime VI is a distribution-of-disease outbreak model calibration code based on variational inference. It accompanies a publication for submission to Statistics in Medicine journal, and the code will be maintained for open-source use on Sandia's GitLab. The software provides methods for calibrating an epidemiological model to measured case-count data for a multitude of correlated spatial regions. The code solves a Bayesian inverse problem for model calibration where the posterior over-model parameters are approximated through a custom implementation of variational inference. Sandia National Laboratories is a multimission laboratory managed and operated by National Technology & Engineering Solutions of Sandia, LLC, a wholly owned subsidiary of Honeywell International Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA0003525.

Safta, Cosmin↗
DOE OSTI2022

A Bayesian Deep Learning Approach to Near-Term Climate Prediction

Since model bias and associated initialization shock are serious shortcomings that reduce prediction skills in state-of-the-art decadal climate prediction efforts, we pursue a complementary machine-learning-based approach to climate prediction. The example problem setting we consider consists of predicting natural variability of the North Atlantic sea surface temperature on the interannual timescale in the pre-industrial control simulation of the Community Earth System Model. While previous works have considered the use of recurrent networks such as convolutional LSTMs and reservoir computing networks in this and other similar problem settings, we currently focus on the use of feedforward convolutional networks. In particular, we find that a feedforward convolutional network with a Densenet architecture is able to outperform a convolutional LSTM in terms of predictive skill. Next, we go on to consider a probabilistic formulation of the same network based on Stein variational gradient descent and find that in addition to providing useful measures of predictive uncertainty, the probabilistic (Bayesian) version improves on its deterministic counterpart in terms of predictive skill. Finally, we characterize the reliability of the ensemble of machine learning models obtained in the probabilistic setting by using analysis tools developed in the context of ensemble numerical weather prediction.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

Wave-function-based emulation for nucleon-nucleon scattering in momentum space

Emulators for low-energy nuclear physics can provide fast and accurate predictions of bound-state and scattering observables for applications that require repeated calculations with different parameters, such as Bayesian uncertainty quantification. In this paper, we extend a scattering emulator based on the Kohn variational principle (KVP) to momentum space (including coupled channels) with arbitrary boundary conditions, which enable the mitigation of spurious singularities known as Kohn anomalies. We test it on a modern chiral nucleon-nucleon (N N) interaction, including emulation of the coupled channels. We provide comparisons between a Lippmann-Schwinger equation emulator and our KVP momentum-space emulator for a representative set of neutron-proton (n p) scattering observables, and also introduce a quasi-spline-based approach for the KVP-based emulator. Furthermore, our findings show that while there are some trade-offs between accuracy and speed, all three emulators perform well. Self-contained Jupyter notebooks that generate the results and figures in this paper are publicly available.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2022

How Bayesian methods can improve R -matrix analyses of data: The example of the d t reaction

The 3 H(d, n) 4 He reaction is of significant interest in nuclear astrophysics and nuclear applications. It is an important, early step in big-bang nucleosynthesis and a key process in nuclear fusion reactors. We use one- and two-level R-matrix approximations to analyze data on the cross section for this reaction at center-of-mass energies below 215 keV. We critically examine the data sets using a Bayesian statistical model that allows for both common-mode and additional point-to-point un- certainties. We use Markov Chain Monte Carlo sampling to evaluate this R-matrix-plus-statistical model and find two-level R-matrix results that are stable with respect to variations in the channel radii. The S factor at 40 keV evaluates to 25.36(19) MeV b (68% credibility interval). We discuss our Bayesian analysis in detail and provide guidance for future applications of Bayesian methods to R-matrix analyses. We also discuss possible paths to further reduction of the S-factor uncertainty.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2022

Adapting Reinforcement Learning Treatment Policies Using Limited Data to Personalize Critical Care

We report reinforcement learning (RL) demonstrates promise for developing effective treatment policies in critical care settings. However, existing RL methods often require large and comprehensive patient data sets and do not readily lend themselves to settings in which certain patient subpopulations are severely underrepresented. In this study, we develop a new method, noisy Bayesian policy updates (NBPU), for selecting high-performing reinforcement learning–based treatment policies for underrepresented patient subpopulations using limited observations. Our method uses variational inference to learn a probability distribution over treatment policies based on a reference patient subpopulation for which sufficient data are available. It then exploits limited data from an underrepresented patient subpopulation to update this probability distribution and adapts its recommendations to this subpopulation. We demonstrate our method’s utility on a data set of ICU patients receiving intravenous blood anticoagulant medication. Our results show that NBPU outperforms state-of-the-art methods in terms of both selecting effective treatment policies for patients with nontypical clinical characteristics and predicting the corresponding policies’ performance for these patients.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2021

OGLE-2018-BLG-1185b: A Low-mass Microlensing Planet Orbiting a Low-mass Dwarf

We report an analysis of the planetary microlensing event OGLE-2018-BLG-1185, which was observed by a large number of ground-based telescopes and by the Spitzer Space Telescope. The ground-based light curve indicates a low planet–host star mass ratio of q = (6.9 ± 0.2) × 10{sup −5}, which is near the peak of the wide-orbit exoplanet mass-ratio distribution. We estimate the host star and planet masses with a Bayesian analysis using the measured angular Einstein radius under the assumption that stars of all masses have an equal probability of hosting the planet. The flux variation observed by Spitzer is marginal, but still places a constraint on the microlens parallax. Imposing a conservative constraint that this flux variation should be Δf {sub Spz} < 4 instrumental flux units yields a host mass of M{sub host}=0.37{sub −0.21}{sup +0.35} M{sub ⊙} and a planet mass of m{sub p}=8.4{sub −4.7}{sup +7.9} M{sub ⊕}. A Bayesian analysis including the full parallax constraint from Spitzer suggests smaller host star and planet masses of M{sub host}=0.091{sub −0.018}{sup +0.064} M{sub ⊙} and m{sub p}=2.1{sub −0.4}{sup +1.5} M{sub ⊕}, respectively. Future high-resolution imaging observations with the Hubble Space Telescope or Extremely Large Telescope could distinguish between these two scenarios and help reveal the planetary system properties in more detail.

47 OTHER INSTRUMENTATION↗