Search NASA⌕ Search

SEARCH · Search NASA

Results for “Bayesian method”

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 685 records · Page 38

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↗

Attributing runoff changes to climate variability and human activities: Uncertainty analysis using four monthly water balance models

Hydrological simulations to delineate the impacts of climate variability and human activities are subjected to uncertainties related to both parameter and structure of the hydrological models. To analyze the impact of these uncertainties on the model performance and to yield more reliable simulation results, a global calibration and multimodel combination method that integrates the Shuffled Complex Evolution Metropolis (SCEM) and Bayesian Model Averaging (BMA) of four monthly water balance models was proposed. The method was applied to the Weihe River Basin (WRB), the largest tributary of the Yellow River, to determine the contribution of climate variability and human activities to runoff changes. The change point, which was used to determine the baseline period (1956-1990) and human-impacted period (1991-2009), was derived using both cumulative curve and Pettitt’s test. Results show that the combination method from SCEM provides more skillful deterministic predictions than the best calibrated individual model, resulting in the smallest uncertainty interval of runoff changes attributed to climate variability and human activities. This combination methodology provides a practical and flexible tool for attribution of runoff changes to climate variability and human activities by hydrological models.

runoff change, uncertainty assessment, monthly wat↗

Dynamic, risk informed decision support systems and methods

The present disclosure is directed to a decision support system or tool based on a Bayesian Network (BN) framework. The diagnostic support tool is created by using advanced Probabilistic Risk Assessment (PRA) method(s) to construct Bayesian Networks (BNs) that form a Bayesian Decision Support Process (BDSP) to provide science-based decision support for understanding and managing events in complex systems. In an embodiment, the PRA method(s) may include Discrete Dynamic Event Trees (DDETs) and simulations.

Groth, Katrina↗

Bayesian High-Rank Hankel Matrix Completion for Nonlinear Synchrophasor Data Recovery

Phasor measurement units (PMUs) provide high temporal-resolution synchrophasor measurements for power system monitoring and control. The frequent data quality issues, such as missing and bad data, prevent the incorporation of synchrophasor data in real-time operations. Most existing data-driven data recovery methods assume the power system dynamics can be approximated by a linear dynamical system, and the recovery performance degrades significantly when the power system is experiencing nonlinear dynamics during significant events. Here, this paper proposes a data-driven Bayesian nonlinear synchrophasor data recovery method (Ba-NSDR) that can recover a consecutive time period of simultaneous data losses or errors across all channels, even when the underlying system is highly nonlinear. The idea is to lift the Hankel matrix of the spatial-temporal synchrophasor data to a higher dimension such that the lifted Hankel matrix is low-rank in that space and can be processed with the kernel trick. Our proposed Bayesian method then infers the probabilistic distributions of synchrophasor from the partial observations. Some distinctive features of Ba-NSDR include an uncertainty index to measure the accuracy of the recovery result and the robustness to parameter selections. Our method is verified on both synthetic and recorded event datasets.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

TH-A-BRF-01: A Probabilistic Bayesian Approach to Derive Electron Density From MRI for Radiation Therapy Treatment Planning

Purpose: MRI significantly improves the accuracy and reliability of target delineation for patient simulation and treatment planning in radiation therapy, due to its superior soft tissue contrast as compared to CT. An MRI based simulation will reduce cost and simplify clinical workflow with zero ionizing radiation. However, MRI lacks the key electron density information. The purpose of this work is to develop a reliable method to derive electron density from MRI. Methods: We adopt a probabilistic Bayesian approach for electron density mapping based on T1-weighted head MRI. For each voxel, we compute conditional probability of electron densities given its: (1) T1 intensity and (2) geometry in a reference anatomy, obtained by deformable image registration between the MRI of test patient and atlas. Intensity and geometry information are combined into a unifying posterior probability density function whose mean gives the electron density. Mean absolute HU error between the estimated and true CT, as well as ROC's for bone detection (HU>200) were calculated for 8 patients. The performance was compared with a global intensity approach based on T1 and no density correction (set whole head to water). Results: The proposed technique significantly reduced the errors in electron density estimation, with a mean absolute HU error of 132, compared with 139 for deformable registration (p=10{sup −3}), 371 for the intensity approach (p=10{sup −5}) and 282 without density correction (p=2×10{sup −4}). For 90% sensitivity in bone detection, the proposed method had a specificity of 85% and that for deformable registration, intensity and without density correction are 80%, 24% and 10% respectively. Conclusion: The proposed unifying method provides accurate electron density estimation and bone detection from MRI of the head with highly heterogeneous regions. This paves the way for accurate dose calculation and generating reference images for patient setup in MRI-based treatment planning.

60 APPLIED LIFE SCIENCES↗

Modeling Soot Oxidation and Gasification with Bayesian Statistics

This paper presents a statistical method for model calibration using data collected from literature. The method is used to calibrate parameters for global models of soot consumption in combustion systems. This consumption is broken into two different submodels: first for oxidation where soot particles are attacked by certain oxidizing agents; second for gasification where soot particles are attacked by H 2 O or CO 2 molecules. Rate data were collected from 19 studies in the literature and evaluated using Bayesian statistics to calibrate the model parameters. Bayesian statistics are valued in their ability to quantify uncertainty in modeling. The calibrated consumption model with quantified uncertainty is presented here along with a discussion of associated implications. The oxidation results are found to be consistent with previous studies. Significant variation is found in the CO 2 gasification rates.

01 COAL, LIGNITE, AND PEAT↗

Bayesian estimation of cross-section and errors of experiment and calculation for accurate prediction of neutronic characteristics

In some experiments, there might be a correlation of experimental errors such as an error caused by fabrication. The correlation of calculation errors also might exist among analyses of experiments. In the present study, considering these correlations, the cross-section and errors of experiment and calculation are adjusted for the accurate prediction of neutronic characteristics. The adjustment method is derived based on Bayesian theory. The 'a posteriori' distributions of cross-section and errors of experiment and calculation are obtained by the derivation. The case study is performed by using the benchmarks of light-water moderated experiments reported in ICSBEP. The error from fuel rod fabrication and measurement are included in the experiment error, and the error from the simplification of the benchmark model is included in the calculation error. Since these errors have a positive correlation among the benchmarks, the prediction accuracy is expected to be improved by considering these correlations. For the sensitivity analysis of the correlation among benchmarks, the correlation coefficient of experiment error and that of calculation error is assumed from 0.0 to 0.9. The results indicate that the prediction accuracy of the multiplication factor can be improved by increasing the correlation coefficient of experiment error and that of calculation error. (authors)

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Technical Report for Bayesian Optimization and Reinforcement Learning for Beam Polarization Increase in the BNL Hadron Injectors

This project developed and evaluated physics-informed Bayesian learning and machine learning (ML)-based optimization methods for improving beam polarization preservation in the BNL hadron injector chain. The work focused on uncertainty-aware digital twin modeling, Bayesian calibration of accelerator simulations using beam measurements, and data-efficient optimization strategies including Bayesian optimization and reinforcement learning. These methods were applied to injector tuning and RF control problems in realistic accelerator settings to support improved operational robustness and readiness for RHIC operations and future Electron–Ion Collider facilities. No subject inventions were disclosed under this award.

43 PARTICLE ACCELERATORS↗

Accelerating Hamiltonian Monte Carlo for Bayesian inference in neural networks and neural operators

Hamiltonian Monte Carlo (HMC) is a powerful and accurate method to sample from the posterior distribution in Bayesian inference. However, HMC techniques are computationally demanding for Bayesian neural networks due to the high dimensionality of the network’s parameter space and the non-convexity of their posterior distributions. Therefore, various approximation techniques, such as variational inference (VI) or stochastic gradient MCMC, are often employed to infer the posterior distribution of the network parameters. Such approximations introduce inaccuracies in the inferred distributions, resulting in unreliable uncertainty estimates. In this work, we propose a hybrid approach that combines inexpensive VI and accurate HMC methods to efficiently and accurately quantify uncertainties in neural networks and neural operators. The proposed approach leverages an initial VI training on the full network. We examine the influence of individual parameters on the prediction uncertainty, which shows that a large proportion of the parameters do not contribute substantially to uncertainty in the network predictions. This information is then used to significantly reduce the dimension of the parameter space, and HMC is performed only for the subset of network parameters that strongly influence prediction uncertainties. This yields a framework for accelerating the full batch HMC for posterior inference in neural networks. We demonstrate the efficiency and accuracy of the proposed framework on deep neural networks and operator networks, showing that inference can be performed for large networks with tens to hundreds of thousands of parameters. Finally, we show that this method can effectively learn surrogates for complex physical systems by modeling the operator that maps from upstream conditions to wall-pressure data on a cone in hypersonic flow.

Bayesian inference↗

Inference, Prediction, & Entropy-Rate Estimation of Continuous-Time, Discrete-Event Processes

Inferring models, predicting the future, and estimating the entropy rate of discrete-time, discrete-event processes is well-worn ground. However, a much broader class of discrete-event processes operates in continuous-time. Here, we provide new methods for inferring, predicting, and estimating them. The methods rely on an extension of Bayesian structural inference that takes advantage of neural network’s universal approximation power. Based on experiments with complex synthetic data, the methods are competitive with the state-of-the-art for prediction and entropy-rate estimation.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

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↗

Evaluating marginal likelihood with thermodynamic integration method and comparison with several other numerical methods

Evaluating marginal likelihood is the most critical and computationally expensive task, when conducting Bayesian model averaging to quantify parametric and model uncertainties. The evaluation is commonly done by using Laplace approximations to evaluate semianalytical expressions of the marginal likelihood or by using Monte Carlo (MC) methods to evaluate arithmetic or harmonic mean of a joint likelihood function. This study introduces a new MC method, i.e., thermodynamic integration, which has not been attempted in environmental modeling. Instead of using samples only from prior parameter space (as in arithmetic mean evaluation) or posterior parameter space (as in harmonic mean evaluation), the thermodynamic integration method uses samples generated gradually from the prior to posterior parameter space. This is done through a path sampling that conducts Markov chain Monte Carlo simulation with different power coefficient values applied to the joint likelihood function. The thermodynamic integration method is evaluated using three analytical functions by comparing the method with two variants of the Laplace approximation method and three MC methods, including the nested sampling method that is recently introduced into environmental modeling. The thermodynamic integration method outperforms the other methods in terms of their accuracy, convergence, and consistency. The thermodynamic integration method is also applied to a synthetic case of groundwater modeling with four alternative models. The application shows that model probabilities obtained using the thermodynamic integration method improves predictive performance of Bayesian model averaging. As a result, the thermodynamic integration method is mathematically rigorous, and its MC implementation is computationally general for a wide range of environmental problems.

97 MATHEMATICS AND COMPUTING↗

ESTIMATING THE PROPERTIES OF HARD X-RAY SOLAR FLARES BY CONSTRAINING MODEL PARAMETERS

We wish to better constrain the properties of solar flares by exploring how parameterized models of solar flares interact with uncertainty estimation methods. We compare four different methods of calculating uncertainty estimates in fitting parameterized models to Ramaty High Energy Solar Spectroscopic Imager X-ray spectra, considering only statistical sources of error. Three of the four methods are based on estimating the scale-size of the minimum in a hypersurface formed by the weighted sum of the squares of the differences between the model fit and the data as a function of the fit parameters, and are implemented as commonly practiced. The fourth method is also based on the difference between the data and the model, but instead uses Bayesian data analysis and Markov chain Monte Carlo (MCMC) techniques to calculate an uncertainty estimate. Two flare spectra are modeled: one from the Geostationary Operational Environmental Satellite X1.3 class flare of 2005 January 19, and the other from the X4.8 flare of 2002 July 23. We find that the four methods give approximately the same uncertainty estimates for the 2005 January 19 spectral fit parameters, but lead to very different uncertainty estimates for the 2002 July 23 spectral fit. This is because each method implements different analyses of the hypersurface, yielding method-dependent results that can differ greatly depending on the shape of the hypersurface. The hypersurface arising from the 2005 January 19 analysis is consistent with a normal distribution; therefore, the assumptions behind the three non-Bayesian uncertainty estimation methods are satisfied and similar estimates are found. The 2002 July 23 analysis shows that the hypersurface is not consistent with a normal distribution, indicating that the assumptions behind the three non-Bayesian uncertainty estimation methods are not satisfied, leading to differing estimates of the uncertainty. We find that the shape of the hypersurface is crucial in understanding the output from each uncertainty estimation technique, and that a crucial factor determining the shape of hypersurface is the location of the low-energy cutoff relative to energies where the thermal emission dominates. The Bayesian/MCMC approach also allows us to provide detailed information on probable values of the low-energy cutoff, E{sub c} , a crucial parameter in defining the energy content of the flare-accelerated electrons. We show that for the 2002 July 23 flare data, there is a 95% probability that E{sub c} lies below approximately 40 keV, and a 68% probability that it lies in the range 7-36 keV. Further, the low-energy cutoff is more likely to be in the range 25-35 keV than in any other 10 keV wide energy range. The low-energy cutoff for the 2005 January 19 flare is more tightly constrained to 107 {+-} 4 keV with 68% probability. Using the Bayesian/MCMC approach, we also estimate for the first time probability density functions for the total number of flare-accelerated electrons and the energy they carry for each flare studied. For the 2002 July 23 event, these probability density functions are asymmetric with long tails orders of magnitude higher than the most probable value, caused by the poorly constrained value of the low-energy cutoff. The most probable electron power is estimated at 10{sup 28.1} erg s{sup -1}, with a 68% credible interval estimated at 10{sup 28.1}-10{sup 29.0} erg s{sup -1}, and a 95% credible interval estimated at 10{sup 28.0}-10{sup 30.2} erg s{sup -1}. For the 2005 January 19 flare spectrum, the probability density functions for the total number of flare-accelerated electrons and their energy are much more symmetric and narrow: the most probable electron power is estimated at 10{sup 27.66{+-}0.01} erg s{sup -1} (68% credible intervals). However, in this case the uncertainty due to systematic sources of error is estimated to dominate the uncertainty due to statistical sources of error.

79 ASTRONOMY AND ASTROPHYSICS↗

Qualitative and quantitative validation of the SINBAD code on complex HPGe gamma-ray spectra

Radionuclides identification and quantification is a serious concern for many applications as safety or security of nuclear power plant or fuel cycle facility, CBRN risk identification, environmental radioprotection and waste measurements. High resolution gamma-ray spectrometry based on HPGe detectors is a performing solution for all these topics. During last decades, a great number of software has been developed to improve gamma spectra analysis. However, some difficulties remain in the analysis when photoelectric peaks are folded together with a high ratio between theirs amplitudes, when the Compton background is much larger compared to the signal of a single peak and when spectra are composed of a great number of peaks. This study deals with the comparison between conventional methods in radionuclides identification and quantification and the code called SINBAD ('Spectrometrie par Inference Non parametrique Bayesienne Deconvolutive'). For many years, SINBAD has been developed by CEA LIST for unfolding complex spectra from HPGe detectors. Contrary to conventional methods using fitting procedures, SINBAD uses a probabilistic approach with Bayesian inference to describe spectrum data. This conventional fitting method founded for example in Genie 2000 is compared with the nonparametric SINBAD approach regarding some key figures of merit as the peak centroid evaluation (identification) and peak surface evaluation (quantification). Unfriendly cases are studied for nuclides detection with closed gamma-rays energies and high photoelectric peak intensity differences. Tests are performed with spectra from the International Atomic Energy Agency (IAEA) for gamma spectra analysis software benchmark and with spectra acquired at the laboratory. Results show that SINBAD and Genie 2000 performances are quite similar with sometimes best results for SINBAD with the important difference that to achieve same performances the nonparametric method is user-friendly compared to the conventional method which need an expert parameters adjustment. (authors)

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

AGNfitter: a Bayesian MCMC approach to fitting spectral energy distributions of AGNs

We present AGNfitter, a publicly available open-source algorithm implementing a fully Bayesian Markov Chain Monte Carlo method to fit the spectral energy distributions (SEDs) of active galactic nuclei (AGNs) from the sub-millimeter to the UV, allowing one to robustly disentangle the physical processes responsible for their emission. AGNfitter makes use of a large library of theoretical, empirical, and semi-empirical models to characterize both the nuclear and host galaxy emission simultaneously. The model consists of four physical emission components: an accretion disk, a torus of AGN heated dust, stellar populations, and cold dust in star-forming regions. AGNfitter determines the posterior distributions of numerous parameters that govern the physics of AGNs with a fully Bayesian treatment of errors and parameter degeneracies, allowing one to infer integrated luminosities, dust attenuation parameters, stellar masses, and star-formation rates. We tested AGNfitter’s performance on real data by fitting the SEDs of a sample of 714 X-ray selected AGNs from the XMM-COSMOS survey, spectroscopically classified as Type1 (unobscured) and Type2 (obscured) AGNs by their optical–UV emission lines. We find that two independent model parameters, namely the reddening of the accretion disk and the column density of the dusty torus, are good proxies for AGN obscuration, allowing us to develop a strategy for classifying AGNs as Type1 or Type2, based solely on an SED-fitting analysis. Our classification scheme is in excellent agreement with the spectroscopic classification, giving a completeness fraction of ∼86% and ∼70%, and an efficiency of ∼80% and ∼77%, for Type1 and Type2 AGNs, respectively.

79 ASTRONOMY AND ASTROPHYSICS↗

Optimal Pulse Processing, Pile-Up Decomposition, and Applications of Silicon Drift Detectors at LCLS

Silicon drift detectors (SDDs) revolutionized spectroscopy in fields as diverse as geology and dentistry. For a subset of experiments at ultrafast, X-ray free-electron lasers (FELs), SDDs can make substantial contributions. Often the unknown spectrum is interesting, carrying science data, or the background measurement is useful to identify unexpected signals. Many measurements involve only several discrete photon energies known a priori, allowing single-event decomposition of pile-up and spectroscopic photon counting. We designed a pulse function and demonstrated that the signal amplitude (i.e., proportional to the detected energy and obtained from fitting with the pulse function), rise time, and pulse height are interrelated, and at short peaking times, the pulse height and pulse area are not optimal estimators for detected energy; instead, the signal amplitude and rise time are obtained for each pulse by fitting, thus removing the need for pulse shaping. By avoiding pulse shaping, rise times of tens of nanoseconds resulted in reduced pulse pile-up and allowed decomposition of remaining pulse pile-up at photon separation times down to hundreds of nanoseconds while yielding time-of-arrival information with the precision of 10 ns. Waveform fitting yields simultaneously high energy resolution and high counting rates (two orders of magnitude higher than current digital pulse processors). At pulsed sources or high photon rates, photon pile-up still occurs. We showed that pile-up spectrum fitting is relatively simple and preferable to pile-up spectrum deconvolution. We then developed a photon pile-up statistical model for constant intensity sources, extended it to variable intensity sources (typical for FELs), and used it to fit a complex pileup spectrum. We subsequently developed a Bayesian pile-up decomposition method that allows decomposing pile-up of single events with up to six photons from six monochromatic lines with 99% accuracy. The usefulness of SDDs will continue into the X-ray FEL era of science. Their successors, the ePixS hybrid pixel detectors, already offer hundreds of pixels, each with a similar performance to an SDD, in a compact, robust and affordable package.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Autonomous organic synthesis for redox flow batteries via flexible batch Bayesian optimization

Traditional trial-and-error methods for materials discovery are inefficient to meet the urgent demands posed by the rapid progression of climate change. This urgency has driven the increasing interest in integrating robotics and machine learning into materials research to accelerate experimental learning. However, idealized decision-making frameworks to achieve maximum sampling efficiency are not always compatible with high-throughput experimental workflows inside a laboratory. For multi-step chemical processes, differences in hardware capacities can complicate the digital framework by introducing constraints on the maximum number of samples in each step of the experiment, hence causing varying batch sizes in variable selection within the same batch. Therefore, designing flexible sampling algorithms is necessary to accommodate the multi-step synthesis with practical constraints unique to each high-throughput workflow. In this work, we designed and employed three strategies on a high-throughput robotic platform to optimize the sulfonation reaction of redox-active molecules used in flow batteries. Our strategies adapt to the multi-step experimental workflow, where their formulation and heating steps are separate, causing varying batch size requirements. By strategically sampling using clustering and mixed-variable batch Bayesian optimization, we were able to iteratively identify optimal conditions that maximize the yields. Our work presents a flexible approach that allows tailoring the machine learning decision-making to suit the practical constraints in individual high-throughput experimental platforms, followed by performing resource-efficient yield optimization using available open-source Python libraries.

Tamura, Clara [Univ. of Washington, Seattle, WA (U↗

Extreme-Scale Bayesian Inference for Uncertainty Quantification of Complex Simulations

Uncertainty quantification (UQ)—that is, quantifying uncertainties in complex mathematical models and their large-scale computational implementations—is widely viewed as one of the outstanding challenges facing the field of CS&E over the coming decade. The EUREKA project set to address the most difficult class of UQ problems: those for which both the underlying PDE model as well as the uncertain parameters are of extreme scale. In the project we worked on these extreme-scale challenges in the following four areas: 1. Scalable parallel algorithms for sampling and characterizing the posterior distribution that exploit the structure of the underlying PDEs and parameter-to-observable map. These include structure-exploiting versions of the randomized maximum likelihood method, which aims to overcome the intractability of employing conventional MCMC methods for solving extreme-scale Bayesian inversion problems by appealing to and adapting ideas from large-scale PDE-constrained optimization, which have been very successful at exploring high-dimensional spaces. 2. Scalable parallel algorithms for construction of prior and likelihood functions based on learning methods and non-parametric density estimation. Constructing problem-specific priors remains a critical challenge in Bayesian inference, and more so in high dimensions. Another challenge is construction of likelihood functions that capture unmodeled couplings between observations and parameters. We will create parallel algorithms for non-parametric density estimation using high dimensional N-body methods and combine them with supervised learning techniques for the construction of priors and likelihood functions. 3. Bayesian inadequacy models, which augment physics models with stochastic models that represent their imperfections. The success of the Bayesian inference framework depends on the ability to represent the uncertainty due to imperfections of the mathematical model of the phenomena of interest. This is a central challenge in UQ, especially for large-scale models. We propose to develop the mathematical tools to address these challenges in the context of extreme-scale problems. 4. Parallel scalable algorithms for Bayesian optimal experimental design (OED). Bayesian inversion yields quantified uncertainties in the model parameters, which can be propagated forward through the model to yield uncertainty in outputs of interest. This opens the way for designing new experiments to reduce the uncertainties in the model parameters and model predictions. Such experimental design problems have been intractable for large-scale problems using conventional methods; we will create OED algorithms that exploit the structure of the PDE model and the parameter-to-output map to overcome these challenges. Parallel algorithms for these four problems were created, analyzed, prototyped, implemented, tuned, and scaled up for leading-edge supercomputers, including UT-Austin’s own 10 petaflops Stampede system, ANL’s Mira system, and ORNL’s Titan system. While our focus is on fundamental mathematical/computational methods and algorithms, we will assess our methods on model problems derived from several DOE mission applications, including multiscale mechanics and ice sheet dynamics.

97 MATHEMATICS AND COMPUTING↗