Search NASA⌕ Search

SEARCH · Search NASA

Results for “Bootstrap”

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 739 records · Page 41

Adaptive AMG with Coarsening Based on Compatible Weighted Matching

We introduce a new composite adaptive Algebraic Multigrid (composite αAMG) method to solve systems of linear equations without a-priori knowledge or assumption on characteristics of near-null components of the AMG preconditioned problem referred to as algebraic smoothness. Our version of αAMG is a composite solver built through a bootstrap strategy aimed to obtain a desired convergence rate. The coarsening process employed to build each new solver component relies on a pairwise aggregation scheme based on maximum weighted matching in a graph and on principles of compatible relaxation. The latter replaces the commonly used characterization of strength of connection in both the coarse space selection and in the interpolation scheme. The goal is to design a method leading to scalable AMG for a wide class of problems that go beyond the standard elliptic Partial Differential Equations (PDEs). In the present work, we introduce the method and demonstrate its potential when applied to symmetric positive definite linear systems arising from finite element discretization of highly anisotropic elliptic PDEs on structured and unstructured meshes.

97 MATHEMATICS AND COMPUTING↗

Experimental Studies of Elementary Particle Interactions at High Energies

This is the final report of a program of research on "Experimental Studies of Elementary Particle Interactions at High Energies'' of the High Energy Physics (HEP) group of The Rockefeller University. The research was carried out using the Collider Detector at Fermilab (CDF) and the Compact Muon Solenoid (CMS) detector at the Large Hadron Collider (LHC) at CERN. Three faculty members, two research associates, and two postdoctoral associates participated in this project. At CDF, we studied proton-antiproton collisions at an energy of 1.96 TeV. We focused on diffractive interactions, in which the colliding antiproton loses a small fraction of its momentum, typically less than 1%, while the proton is excited into a high mass state retaining its quantum numbers. The study of such collisions provides insight into the nature of the diffractive exchange, conventionally referred to as Pomeron exchange. In studies of W and Z production, we found results that point to a QCD-based interpretation of the diffractive exchange, as predicted in a data-driven phenomenology developed within the Rockefeller HEP group. At CMS, we worked on diffraction, supersymmetry (SUSY), dark matter, large extra dimensions, and statistical applications to data analysis projects. In diffraction, we extended our CDF studies to higher energies working on two fronts: measurement of the single/double diffraction and of the rapidity gap cross sections at 7 TeV, and development of a simulation of diffractive processes along the lines of our successful model used at CDF. Working with the PYTHIA8 Monte Carlo simulation authors, we implemented our model as a PYTHIA8-MBR option in PYTHIA8 and used it in our data analysis. Preliminary results indicate good agreement. We searched for SUSY by measuring parameters in the Constrained Minimal Supersymmetric extension of the Standard Model (CMSSM) and found results which, combined with other experimental constraints and theoretical considerations, indicate that the CMSSM is not a viable model. Expressing our results in terms of simple topologies, we exclude squark masses below 0.75 TeV and gluino masses below 1.1 TeV. Astrophysical measurements suggest that about 80% of the matter density of the Universe is non-luminous. One of the theories on dark matter attributes it to Weakly Interacting Massive Particles (WIMPs). We searched for WIMPs in 7 TeV and 8 TeV collisions at CMS and set limits on WIMP production rates, which are competitive and complementary to those of direct detection experiments. Searching for monojets (events with only one jet), which in a popular model could be produced by a jet paired by a gravitino that escapes into extra dimensions, we significantly improved the previously set limit. Our results have been used to set limits on Higgs decay to invisible particles and on production of top squarks in compressed SUSY scenarios. Statistics. We computed Bayesian reference priors for several types of measurement and used them in the analysis of CMS data; investigated the applicability of bootstrap methods to HEP measurements; studied several issues associated with simple-versus-simple hypothesis testing and applied the resulting methods to the measurement of some properties of the top quark and Higgs boson.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

The Physics of Basis For A Conservative Physics And Conservative Technology Tokamak Power Plant, ARIES-ACT2

The conservative physics and conservative technology tokamak power plant ARIES-ACT2 has a major radius of 9.75 m at aspect ratio of 4.0, strong shaping with elongation of 2.2 and triangularity of 0.63. The no wall βN reaches ~ 2.4, limited by n=1 external kink mode, and can be extended to 3.2 with a stabilizing shell behind the ring structure shield. The bootstrap current fraction is 77% with a q95 of 8.0, requiring about ~ 4.0 MA of external current drive. This current is supplied with 30 MW of ICRF/FW and 80 MW of negative ion NB. Up to 1.0 MA can be driven with LH with no wall, and 1.5 or more MA can be driven with a stabilizing shell. EC was examined and is most effective for safety factor control over ρ ~ 0.2-0.6 with 20 MW. The pedestal density is ~ 0.65x10 20 /m 3 and the temperature is ~ 9.0 keV. The H98 factor is 1.25, n/n Gr = 1.3, and the net power to LH threshold power is 1.3-1.4 in the flattop. Due to the high toroidal field and high central temperature the cyclotron radiation loss was found to be high depending on the first wall reflectivity.

(MHD) RF Heating↗

The Physics Basis For An Advanced Physics And Advanced Technology Tokamak Power Plant Configuration, ARIES-ACT1

The advanced physics and advanced technology tokamak power plant ARIES-ACT1 has a major radius of 6.25 m at aspect ratio of 4.0, toroidal field of 6.0 T, strong shaping with elongation of 2.2 and triangularity of 0.63. The broadest pressure cases reached wall stabilized β N ~ 5.75, limited by n=3 external kink mode requiring a conducting shell at b/a = 0.3, and requiring plasma rotation, feedback, and or kinetic stabilization. The medium pressure peaking case reached βN = 5.28 with BT = 6.75, while the peaked pressure case reaches β N < 5.15. Fast particle MHD stability shows that the alpha particles are unstable, but this leads to redistribution to larger minor radius rather than loss from the plasma. Edge and divertor plasma modeling show that about 75% of the power to the divertor can be radiated with an ITER-like divertor geometry, while over 95% can be radiated in a stable detached mode with an orthogonal target and wide slot geometry. The bootstrap current fraction is 91% with a q 95 of 4.5, requiring about ~ 1.1 MA of external current drive. This current is supplied with 5 MW of ICRF/FW and 40 MW of LHCD. EC was examined and is most effective for safety factor control over ρ ~ 0.2-0.6 with 20 MW. The pedestal density is ~ 0.9x10 20 /m 3 and the temperature is ~ 4.4 keV. The H 98 factor is 1.65, n/n Gr = 1.0, and the net power to LH threshold power is 2.8-3.0 in the flattop.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Detecting and Testing for Structural Error in Computer Models with Application to the Community Atmospheric Model

Bayesian calibration/inversion methods for uncertain input parameters to simulation codes have generated an increasing amount of interest in the climate community in the recent years. This approach uses multiple disparate observational datasets related to the output of the simulator of interest to yield a probabilistic constraint on the uncertain inputs (i.e. probability distribution). A major challenge in applying Bayesian calibration methods to simulation codes is related to the possible presence of a large structural error in the output fields of interest; that is, the presence of discrepancy between the observed and the simulated output that is neither explained by the uncertainty in the inputs nor the observations. We present here methods to test for and visualize structural error in an ensemble of input-perturbed simulations. The methods are based on bootstrap (resmpling) test statistics and well-known metrics to quantify the difference between two probability distributions. The proposed methods are demonstrated using selected precipitation output fields of the Community Atmospheric Model (CAM), which are compared to observations provided by the Global Precipitation Climatology Project (GPCP). The methods are found to be useful in identifying regions and seasons with a large structural error. The proposed structural error detection methodology can be used to screen multiple observational datasets of interest for structural error, making it possible to retain only the datasets with a small structural error for use in Bayesian calibration. In addition, information about the amount of structural error can be valuable to model developers.

58 GEOSCIENCES↗

Unit Price Scaling Trends for Chemical Products

To facilitate early-stage life-cycle techno-economic modeling of emerging technologies, here we identify scaling relations between unit price and sales quantity for a variety of chemical products of three categories - metal salts, organic compounds, and solvents. We collect price quotations for lab-scale and bulk purchases of chemicals from both U.S. and Chinese suppliers. We apply a log-log linear regression model to estimate the price discount effect. Using the median discount factor of each category, one can infer bulk prices of products for which only lab-scale prices are available. We conduct out-of-sample tests showing that most of the price proxies deviate from their actual reference prices by a factor less than ten. We also apply the bootstrap method to determine if a sample median discount factor should be accepted for price approximation. We find that appropriate discount factors for metal salts and for solvents are both -0.56, while that for organic compounds is -0.67 and is less representative due to greater extent of product heterogeneity within this category.

97 MATHEMATICS AND COMPUTING↗

Uncertainty Quantification for Nuclear Safeguards and Non-Destructive Assay - Fact Sheet

UQ is the scientific art of generating confidence statements. Without defensible UQ physical measurements and calculations have no meaning. The UQ team performed a series of case studies with the following results: minimum detectable activity of a Tomographic Gamma-Ray Scanning system was determined using the Currie formalism. Bootstrapping method was used to generate fluorescence yield parameters and uncertainties. For Hybrid K-Edge Densitometry (HKED) measurements, these parameters can be used to predict the concentration of plutonium and quantify uncertainty in HKED models for nuclear safeguards measurements. Detection efficiency was determined for a coincidence counter using covariance data. The use of covariance information drastically reduced the total uncertainty in the average detection efficiency.

98 NUCLEAR DISARMAMENT, SAFEGUARDS, AND PHYSICAL P↗

AGR-1, AGR-2 and AGR-3/4 Dimensional Change Data Analysis

A series of Advanced Gas Reactor (AGR) experiments have been completed in the Advanced Test Reactor at Idaho National Laboratory in support of qualification and development of tristructural isotropic fuel. Each AGR test consists of multiple independently controlled and monitored capsules containing fuel compacts placed in a graphite cylinder. These capsules are instrumented with thermocouples embedded in the graphite, enabling temperature control. The fuel compacts are composed of fuel particles surrounded by a graphitic A3 matrix material. Dimensional change in AGR fuel compacts is vital because the swelling or shrinkage affects the size of the gas gaps that are used to control temperatures. Analysis of dimensional change in the AGR fuel compacts is needed to establish the variables directly relating to compact shrinkage. The variables initially identified for consideration were matrix density, compact density, fuel packing fraction, uranium loading, fuel particle diameter, cumulative fast neutron fluence, and volume average time average fuel temperature. In addition to the data available from the AGR experiments, the analysis included specimens formed from the same A3 matrix material used in Advanced Graphite Creep (AGC) experiments, which provide graphite creep data during irradiation for design and licensing purposes. The primary purpose of including the AGC specimens was to encompass dimensional behavior at zero packing fraction, zero uranium loading, and zero particle diameter. All possible combinations of first-order variable regressions were considered in the analysis. The study focused on identifying the best regression models for percent change in diameter, length, and volume. Bootstrap analysis was used to ensure the resulting regression models were robust and well-performing. The variables identified as very significant in predicting change in one or more dimensions (diameter, length, and volume) are volume average time average temperature, fast fluence, compact density, packing fraction and fuel particle diameter. Due to the presence of confounding effects between several variables, interpretation of these results is equivocal; the use of multiple statistical tests provides additional confidence in the conclusion.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Proposal 1114.11.2956B

The Theoretical Advanced Study Institute was held at the University of Colorado, Boulder, during June 1 - 26, 2015. The topic was "New Frontiers in Fields and Strings." Topics included many discussions of entanglement entropy, the conformal bootstrap, AdS/CFT techniques and applications, cosmology, and the black hole information problem. The organizers were Professors Joseph Polchinski (KITP Santa Barbara) and Pedro Vieira (Perimeter Institute). Sixty-one students heard sixty-two lectures by sixteen lecturers. A Proceedings is in press.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Uncertainty Quantification and Comparison of Weld Residual Stress Measurements and Predictions.

In pressurized water reactors, the prevention, detection, and repair of cracks within dissimilar metal welds is essential to ensure proper plant functionality and safety. Weld residual stresses, which are difficult to model and cannot be directly measured, contribute to the formation and growth of cracks due to primary water stress corrosion cracking. Additionally, the uncertainty in weld residual stress measurements and modeling predictions is not well understood, further complicating the prediction of crack evolution. The purpose of this document is to develop methodology to quantify the uncertainty associated with weld residual stress that can be applied to modeling predictions and experimental measurements. Ultimately, the results can be used to assess the current state of uncertainty and to build confidence in both modeling and experimental procedures. The methodology consists of statistically modeling the variation in the weld residual stress profiles using functional data analysis techniques. Uncertainty is quantified using statistical bounds (e.g. confidence and tolerance bounds) constructed with a semi-parametric bootstrap procedure. Such bounds describe the range in which quantities of interest, such as means, are expected to lie as evidenced by the data. The methodology is extended to provide direct comparisons between experimental measurements and modeling predictions by constructing statistical confidence bounds for the average difference between the two quantities. The statistical bounds on the average difference can be used to assess the level of agreement between measurements and predictions. The methodology is applied to experimental measurements of residual stress obtained using two strain relief measurement methods and predictions from seven finite element models developed by different organizations during a round robin study.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Summary of Documentation for DYNA3D-ParaDyn's Software Quality Assurance Regression Test Problems

The Software Quality Assurance (SQA) regression test suite for DYNA3D (Zywicz and Lin, 2015) and ParaDyn (DeGroot, et al., 2015) currently contains approximately 600 problems divided into 21 suites, and is a required component of ParaDyn’s SQA plan (Ferencz and Oliver, 2013). The regression suite allows developers to ensure that software modifications do not unintentionally alter the code response. The entire regression suite is run prior to permanently incorporating any software modification or addition. When code modifications alter test problem results, the specific cause must be determined and fully understood before the software changes and revised test answers can be incorporated. The regression suite is executed on LLNL platforms using a Python script and an associated data file. The user specifies the DYNA3D or ParaDyn executable, number of processors to use, test problems to run, and other options to the script. The data file details how each problem and its answer extraction scripts are executed. For each problem in the regression suite there exists an input deck, an eight-processor partition file, an answer file, and various extraction scripts. These scripts assemble a temporary answer file in a specific format from the simulation results. The temporary and stored answer files are compared to a specific level of numerical precision, and when differences are detected the test problem is flagged as failed. Presently, numerical results are stored and compared to 16 digits. At this accuracy level different processor types, compilers, number of partitions, etc. impact the results to various degrees. Thus, for consistency purposes the regression suite is run with ParaDyn using 8 processors on machines with a specific processor type (currently the Intel Xeon E5530 processor). For non-parallel regression problems, i.e., the two XFEM problems, DYNA3D is used instead. When environments or platforms change, executables using the current source code and the new resource are created and the regression suite is run. If differences in answers arise, the new answers are retained provided that the differences are inconsequential. This bootstrap approach allows the test suite answers to evolve in a controlled manner with a high level of confidence. Developers also run the entire regression suite with (serial) DYNA3D. While these results normally differ from the stored (parallel) answers, abnormal termination or wildly different values are strong indicators of potential issues.

42 ENGINEERING↗

Stochastic Richardson Extrapolation Based Numerical Error Estimation for Kinetic Plasma Simulations

We present a numerical error estimation technique specifically formulated to deal with stochastic code output with multiple discretization parameters. This method is based on multiple fits to an error model with arbitrary convergence rates and cross-coupling terms, performed using nonlinear optimization. The fitting approach varies by the type of residual norm which influences the importance of outliers, and weights which influences the relative importance of data points in the coarse and refined regions of discretization parameter space. To account for the influence of stochastic noise, these fits are performed on multiple bootstrap values based on the underlying response data set. Using an automated discretization domain selection scheme, the fits are performed on a series of reduced sets of discretization levels in order to find an optimal fully-converged result estimate in the minimum variance sense; this automated approach enables straightforward analysis of multiple quantities of interest and/or time and spatially-dependent response data. The overall numerical error analysis method is useful for verification and validation problems for stochastic simulation methods and forms a key component in the overall uncertainty quantification process. The method was demonstrated for steady and unsteady electron diode problems simulated using a particle-in-cell kinetic plasma code, demonstrating excellent results.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Subcontract B629176: Algebraic Multigrid with Optimal Interpolation (Final Report)

Progress has been made in several of the proposed areas of research, as summarized below: Compatible Relaxation and Optimal Interpolation and Bootstrap Algebraic Multigrid and non-Galerkin AMG for the Dirac equations. Our preliminary findings are promising and suggest that these techniques should lead to efficient and robust solvers for various discretizations of the Dirac PDE.

97 MATHEMATICS AND COMPUTING↗

Collaboration with NSTX in Calculations of Radiofrequency and Neutral Beam Heating and Current Drive

The overall objective of this project was to advance the understanding of heating and current drive necessary for toroidal magnetic confinement fusion energy reactors, and to provide accurate models for these sources to aid in understanding magneto-hydrodynamic, transport and boundary physics issues. Increased understanding is provided by computer simulation using the underlying physics laws, and will lead to greater reliability and increased efficiency of a power producing fusion reactor. The proposed work entailed applications of existing CompX Fokker-Planck, ray tracing, full-wave microwave, and particle diffusion calculator codes to the National Spherical Torus Experiment at Princeton Plasma Physics Laboratory. Special features of NSTX include that large ion orbits are emphasized in its geometry and thus offer a major opportunity within fusion research for focused exploration of their ubiquitous effects, including comparison with simulation codes. The CompX CQL3D Fokker-Planck code was be further adapted to the finite-orbit-width physics. Comparison with experiment ensured that the computer codes have sufficient accuracy to accurately model the experiments. Nonthermal distributions of energetic ions are produced in experiments-- and modeled with the codes-- exhibiting the signatures of auxiliary heating, transport, and radiation processes within the plasma. Using physics-based modeling, in conjunction with experimental data, nonthermal distributions can provide keys to validation and quantification of auxiliary heating physics. Nonthermal distributions also hold particular information on radial transport processes, not obscured by the smoothing Maxwellization effects of Coulomb collisions. Generally, radial transport is not well understood but is a key to efficient fusion energy production. By comparison with Fokker-Planck modeling, the velocity dependence of radial transport coefficients becomes experimentally accessible, particularly in cases of well-modeled sources such as neutral beams, fast wave power, or the toroidal electric field. It is clear from modeling of fast-wave interactions with ions that finite-orbit-width effects play a significant role. For example, a different set of ions is computed to intersect a given cyclotron resonance depending on whether or not radial orbit drifts are taken into account. Also, radial transport, associated bootstrap current, and alpha channeling depend on radial drifts. Ion cyclotron heating above the first harmonic interacts preferentially with higher velocity ions, giving large radial drifts. The CQL3D Fokker-Planck code, maintained by the proposers, is a physics based model used extensively within the US fusion community to compare with and interpret experimental observations. At the beginning of this grant, bounce-averages of the rf quasilinear and other operators have been performed neglecting the radial drifts which give finite orbit widths to the particle orbits. The proposed work included a major upgrade to CQL3D to include finite-orbit-width effects be accounted for by performing the bounce-averages of rf and neutral beam fast ions along guiding center orbits. This CQL3D enhancement provided microwave and neutral beam contributions to neoclassical radial transport coefficients for nonthermal distributions, and substantially increased the comprehensiveness of the ion model. Additional work validated Fokker-Planck and ray tracing modeling of neutral beam and high harmonic fast wave absorption and current drive against the NSTX experiment, as had been very successful for electron cyclotron current drive in the DIII-D tokamak at General Atomics and lower hybrid waves on C-Mod at MIT. The validation work was in conjunction with staff and graduate students at Princeton Plasma Physics Laboratory. This DOE grant supported fusion energy research, a potential long-term solution to the world's energy needs. The research grant under consideration, DE-SC0006614, provided research support at the 0.3 Full-Time-Equivalent level for application of computer models to aid in understanding and projecting efficacy of heating and current drive sources in the National Spherical Torus Experiment, a tokamak variant, at the Princeton Plasma Physics Laboratory. The NSTX experiment explores the physics of very tight aspect ratio, almost spherical tokamaks, aiming at producing steady-state fusion plasmas. The current drive is an integral part of the steady-state concept, maintaining the magnetic geometry in the steady-state tokamak.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Energetic particle transport in 3D equilibrium. Final report

In this project, we have further developed the state-of-the-art fusion simulation code, gyrokinetic toroidal code (GTC), for studying the energetic particle (EP) transport in the 3D equilibrium. We have applied the new capabilities to simulate the EP transport by the Alfven eigenmodes in stellarators and by 3D magnetic islands in tokamaks. As part of V&V efforts for 3D capability, the 3D capabilities of GTC have been applied to study thermal transport in stellarators and tokamak with 3D magnetic fields by microturbulence, which could cause additional EP loss. Effects magnetic islands on neoclassical bootstrap current have also been studied. The computational capability developed and physics learned in this project are in support of fusion SciDAC Center for Integrated Simulation of Energetic Particles in Burning Plasmas (ISEP). During this project, the PI has supervised 6 postdoctoral researchers and 4 PhD students at the University of California, Irvine. A complete list of 49 publications and 23 invited talks partially supported by this grant since 7/2015 appears at the end of this report. Five key publications closely related to this project are highlighted in this report.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Conservative Estimation of Tail Probabilities from Limited Sample Data

Several sparse-sample uncertainty quantification (UQ) methods are compared for conservative but not overly conservative estimation of small tail probabilities involving responses that lay beyond specified thresholds in the tails of probability distributions. Sixteen very differently shaped distributions (or probability density functions, PDFs) and tail probability magnitudes ranging from 10 -5 to 10 -1 are considered in order for the study to be relevant to a wide range of risk analysis and quantification of margins and uncertainty (QMU) problems. The emphasis of the study is on limited data regimes ranging from N = 2 to 20 samples, reflective of most experimental and some expensive computational situations. Relatively simple sparse-sample UQ methods tested for this regime involve statistical tolerance interval "Equivalent Normal and related "Ensemble of Normals" and "Superdistribution (SD) approaches. (The independently derived SD is effectively equivalent to the Bayesian posterior predictive distribution given the assumptions of the derivation.) The performance of the methods was generally improved for N ≥ 5 samples with a generalized Jackknife resampling technique, which determines a tail probability estimate by averaging estimates from smaller sub-samples. Several quantitative metrics for method conservatism and accuracy of tail probability estimation are used to assess and rank the methods' performance over many random trials for each test PDF and probability magnitude. A variant of Bootstrap resampling was also tried, but did not significantly improve tail probability estimates in most cases. Detailed results are presented from over 100-million tests over the above factors that provide useful granular information on which methods or combination of methods perform best in various areas of the factor space.

97 MATHEMATICS AND COMPUTING↗

Multiclass classification experiments

We seek a multiclass classifier that can satisfy the following requirements. 1. Predict among at least three classes. 2. Handle a moderately large number of correlated predictors. 3. Handle mixed categorical and continuous predictors. 4. Handle missing values (in some way). 5. Be trained with about n = 1,000 responses, and 6. Quantify prediction uncertainty. Based on 5. above, it is assumed that accuracy of prediction is paramount (relative to, say, runtime). All methods considered in this document take on the order of minutes (most take seconds) on a 2.8 GHz Quad-Core Intel Core i7 processor with 16GB RAM. Secondly, quantification of uncertainty is assumed to be a secondary task, since, with this small data size, resampling methods (i.e. bootstrapping) are possible. Further studies should quantify the feasibility and accuracy of resampling methods, and/or other methods, for quantifying uncertainty. R was used for all software.

97 MATHEMATICS AND COMPUTING↗

Bayesian Model Calibration for Extrapolative Prediction via Gibbs Posteriors

The current standard Bayesian approach to model calibration, which assigns a Gaussian process prior to the discrepancy term, often suffers from issues of unidentifiability and computational complexity and instability. When the goal is to quantify uncertainty in physical parameters for extrapolative prediction, then there is no need to perform inference on the discrepancy term. With this in mind, we introduce Gibbs posteriors as an alternative Bayesian method for model calibration, which updates the prior with a loss function connecting the data to the parameter. The target of inference is the physical parameter value which minimizes the expected loss. We propose to tune the loss scale of the Gibbs posterior to maintain nominal frequentist coverage under assumptions of the form of model discrepancy, and present a bootstrap implementation for approximating coverage rates. Our approach is highly modular, allowing an analyst to easily encode a wide variety of such assumptions. Furthermore, we provide a principled method of combining posteriors calculated from data subsets. We apply our methods to data from an experiment measuring the material properties of tantalum.

36 MATERIALS SCIENCE↗