Search NASA⌕ Search

SEARCH · Search NASA

Results for “Marginal likelihood”

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 19 records
All sourcesNASA NTRSDOE OSTI
DOE OSTI2026

Non -degenerate marginal-likelihood calibration with application to quantum characterization

Here, we propose a marginal likelihood strategy within the Kennedy-O’Hagan (KOH) Bayesian framework, where a Gaussian process (GP) models the discrepancy between a physical system and its simulator. Our approach introduces a novel marginalized likelihood by integrating out the degenerate eigenspace of the covariance matrix, rather than approximating the original likelihood. Unlike approximation methods that compromise accuracy for computational efficiency, our method defines an exact likelihood—distinct from the original but preserving all relevant information. This formulation achieves computational efficiency and stability, even for large datasets where the covariance matrix nears degeneracy. Applied to the characterization of a superconducting quantum device at Lawrence Livermore National Laboratory, the approach enhances the predictive accuracy of the Lindblad master equations for modeling Ramsey measurement data by effectively quantifying uncertainties consistent with the quantum data.

general physics↗
DOE OSTI2021

Development and application of marginal likelihood optimization for integral parameter adjustment

When adjusting nuclear data with integral experiments, care must be taken that spurious adjustments are not made by assimilating poorly characterized integral parameters. If there are unaccounted for biases or poorly estimated uncertainties in the calculated and experimental values for an integral parameter, the Bayesian data assimilation may adjust the nuclear data in a manner that does not reflect the physics of the integral parameter. To identify and lessen the impact of these inconsistent integral parameters, in this study we present a Marginal Likelihood Optimization algorithm. In a data-driven way, the marginalized likelihood is used to modulate hyperparameter terms that decrease the influence of inconsistent integral parameters on the adjustment. The advantage of this approach over other methods in the literature is that it incorporates correlation information and does not remove an integral parameter from the adjustment. Herein, we present and motivate the algorithm, and apply it to an integral data assimilation case study.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
NASA NTRSDate not supplied

Sampling Functions from Gaussian Processes and Structured Covariance Gaussian Networks

When learning aerodynamic models from data, it is critical to incorporate estimates of model uncertainty. This motivates the design of probabilistic aerodynamic databases which can be sampled to generate physically and statistically plausible aerodynamic models. In this talk we discuss how to sample deterministic functions from two different kinds of probabilistic models and demonstrate their use. First, Gaussian Process Regressors (GPRs) are a widely used probabilistic kernel-based model which can be thought of as Gaussian distributions over functions. GPRs are generally trained by maximizing the marginal likelihood of seeing the training data over the kernel parameter space. Sample functions are easily generated by drawing points from the Gaussian distribution at desired input points. However, when the points are not known ahead of time, the classical sampling approach is not possible since successive function samples will generate different function realizations. We present an approach for sampling consistent function evaluations from a GPR over multiple samples. Second, we describe a neural network architecture which learns a conditional Gaussian distribution by maximizing the marginal likelihood at each point in the input space. We then discuss and compare several options for generating sample functions which match this distribution. Finally, we demonstrate the use of these probabilistic aerodynamic models in an atmospheric reentry simulation.

Gaussian process regression↗
NASA NTRS1994

Cosmic temperature fluctuations from two years of COBE differential microwave radiometers observations

The first two years of Cosmic Background Explorer (COBE) Differential Microwave Radiometers (DMR) observations of the cosmic microwave background (CMB) anisotropy are analyzed and compared with our previously published first year results. The results are consistent, but the addition of the second year of data increases the precision and accuracy detected CMB temperature fluctuations. The 2 yr 53 GHz data are characterized by rms temperature fluctuations of (delta-T)(sub rms) (7 deg) = 44 +/- 7 micro-K and (delta-T)(sub rms) (10 deg) = 30.5 +/- 2.7 micro-K at 7 deg and 10 deg angular resolution, respectively. The 53 x 90 GHz cross-correlation amplitude at zero lag is C(0)(sup 1/2) = 36 +/- 5 micro-K (68% CL) for the unsmoothed (7 deg resolution) DMR data. We perform a likelihood analysis of the cross-correlation function, with Monte Carlo simulations to infer biases of the method, for a power-law model of initial density fluctuations, P(k) proportional to R(exp n). The Monte Carlo simulations indicate that derived estimates of n are biased by +0.11 +/- 0.01, while the subset of simulations with a low quadrupole (as observed) indicate a bias of +0.31+/- 0.04. Derived values for 68% confidence intervals are given corrected (and not corrected) for our estimated biases. Including the quadrupole anisotropy, the most likely quadrupole-normalized amplitude is Q(sub rms-PS) = 14.3(sup + 5.2 sub -3.3) micro-K (12.8(sup + 5.2 sub -3.3) micro-K0 with a spectral index n = 1.42(sup + 0.49 sub -0.55)(n = 1.53(sup + 0.49 sub -0.55). With n fixed to 1.0 the most likely amplitude is 18.2 +/- 11.5 micro-K (17.4 +/- 1.5 micro-K). The marginal likelihood of n is 1.42 +/- 0.37 (1.53 +/- 0.37). Excluding the quadrupole anisotropy, the most likely quadrupole-normalized amplitude is Q(sub rms-PS) = 17.4(sup + 7.5 sub -5.2) micro-K (15.8(sup + 7.5 sub -5.2) micro-K) with a spectral index n = 1.11(sup + 0.60 sub -0.55) (n = 1.22(sup + 0.60 sub -0.55). With n fixed to 1.0 the most likely amplitude is 18.6 +/- 1.6 micro-K (18.2 +/- 1.6 micro-K). The marginal likelihood of n is 1.11 +/- 0.40 (1.22 +/- 0.40). Our best estimate of the dipole from the 2 yr DMR data is 3.363 +/- 0.024 mK toward Galactic coordinates (l, b) = (264.4 deg +/- 0.2 deg, 48.1 deg +/- 0.4 deg), and our best estimate of the rms quadrupole amplitude in our sky is 6 +/- 3 micro-K (68% CL).

Bennett, C. L.↗
DOE OSTI2020

Modeling strength and failure variability due to porosity in additively manufactured metals

To model and quantify the variability in plasticity and failure of additively manufactured metals due to imperfections in their microstructure, we have developed uncertainty quantification methodology based on pseudo marginal likelihood and embedded variability techniques. We account for both the porosity resolvable in computed tomography scans of the initial material and the sub-threshold distribution of voids through a physically motivated model. Calibration of the model indicates that the sub-threshold population of defects dominates the yield and failure response. Finally, the technique also allows us to quantify the distribution of material parameters connected to microstructural variability created by the manufacturing process, and, thereby, make assessments of material quality and process control.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Persistent Sampling: Enhancing the Efficiency of Sequential Monte Carlo

Sequential Monte Carlo (SMC) samplers are powerful tools for Bayesian inference but suffer from high computational costs due to their reliance on large particle ensembles for accurate estimates. We introduce persistent sampling (PS), an extension of SMC that systematically retains and reuses particles from all prior iterations to construct a growing, weighted ensemble. By leveraging multiple importance sampling and resampling from a mixture of historical distributions, PS mitigates the need for excessively large particle counts, directly addressing key limitations of SMC such as particle impoverishment and mode collapse. Crucially, PS achieves this without additional likelihood evaluations-weights for persistent particles are computed using cached likelihood values. This framework not only yields more accurate posterior approximations but also produces marginal likelihood estimates with significantly lower variance, enhancing reliability in model comparison. Furthermore, the persistent ensemble enables efficient adaptation of transition kernels by leveraging a larger, decorrelated particle pool. Experiments on high-dimensional Gaussian mixtures, hierarchical models, and non-convex targets demonstrate that PS consistently outperforms standard SMC and related variants, including recycled and waste-free SMC, achieving substantial reductions in mean squared error for posterior expectations and evidence estimates, all at reduced computational cost. PS thus establishes itself as a robust, scalable, and efficient alternative for complex Bayesian inference tasks.

Karamanis, Minas↗
DOE OSTI2021

Dark Matter Constraints from a Unified Analysis of Strong Gravitational Lenses and Milky Way Satellite Galaxies

Joint analyses of small-scale cosmological structure probes have been relatively unexplored yet promise to advance measurements of microphysical dark matter properties using heterogeneous data. Here, we present a multidimensional analysis of dark matter substructure using strong gravitational lenses and the Milky Way (MW) satellite galaxy population, accounting for degeneracies in model predictions and using covariances in the constraining power of these individual probes for the first time. We simultaneously infer the projected subhalo number density and the half-mode mass describing the suppression of the subhalo mass function in thermal relic warm dark matter (WDM), $M_{\mathrm{hm}}$, using the semianalytic model \texttt{Galacticus} to connect the subhalo population inferred from MW satellite observations to the strong lensing host halo mass and redshift regime. Combining MW satellite and strong lensing posteriors in this parameter space yields $M_{\mathrm{hm}}<10^{7.0}\ M_{\mathrm{\odot}}$ (WDM particle mass $m_{\mathrm{WDM}}>9.7\ \mathrm{keV}$) at $95\%$ confidence and disfavors $M_{\mathrm{hm}}=10^{7.4}\ M_{\mathrm{\odot}}$ ($m_{\mathrm{WDM}}=7.4\ \mathrm{keV}$) with a 20:1 marginal likelihood ratio, improving limits on $m_{\mathrm{WDM}}$ set by the two methods independently by $\sim 30\%$. These results are marginalized over the line-of-sight contribution to the strong lensing signal, the mass of the MW host halo, and the efficiency of subhalo disruption due to baryons and are robust to differences in the disruption efficiency between the MW and strong lensing regimes at the $\sim 10\%$ level. This work paves the way for unified analyses of next-generation small-scale structure measurements covering a wide range of scales and redshifts.

79 ASTRONOMY AND ASTROPHYSICS↗
NASA NTRS2021

A NICER View of the Massive Pulsar PSR J0740+6620 Informed by Radio Timing and XMM-Newton Spectroscopy

We report on Bayesian estimation of the radius, mass, and hot surface regions of the massive millisecond pulsar PSR J0740+6620, conditional on pulse-profile modeling of Neutron Star Interior Composition Explorer X-ray Timing Instrument event data. We condition on informative pulsar mass, distance, and orbital inclination priors derived from the joint North American Nanohertz Observatory for Gravitational Waves and Canadian Hydrogen Intensity Mapping Experiment/Pulsar wideband radio timing measurements of Fonseca et al. We use XMM-Newton European Photon Imaging Camera spectroscopic event data to inform our X-ray likelihood function. The prior support of the pulsar radius is truncated at 16 km to ensure coverage of current dense matter models. We assume conservative priors on instrument calibration uncertainty. We constrain the equatorial radius and mass of PSR J0740+6620 to be-+12.390.981.30km and-+2.0720.0660.067Me respectively, each reported as the posterior credible interval bounded by the 16% and 84% quantiles, conditional on surface hot regions that are non-overlapping spherical caps of fully ionized hydrogen atmosphere with uniform effective temperature; a posteriori, the temperature is=-+TlogK5.99100.060.05([])for each hot region. All software for the X-ray modeling framework is open-source and all data, model, and sample information is publicly available, including analysis notebooks and model modules in the Python language. Our marginal likelihood function of mass and equatorial radius is proportional to the marginal joint posterior density of those parameters(within the prior support)and can thus be computed from the posterior samples.

Millisecond pulsars↗
DOE OSTI2021

A NICER View of the Massive Pulsar PSR J0740+6620 Informed by Radio Timing and XMM-Newton Spectroscopy

We report on Bayesian estimation of the radius, mass, and hot surface regions of the massive millisecond pulsar PSR J0740+6620, conditional on pulse-profile modeling of Neutron Star Interior Composition Explorer X-ray Timing Instrument event data. We condition on informative pulsar mass, distance, and orbital inclination priors derived from the joint North American Nanohertz Observatory for Gravitational Waves and Canadian Hydrogen Intensity Mapping Experiment/Pulsar wideband radio timing measurements of Fonseca et al. We use XMM-Newton European Photon Imaging Camera spectroscopic event data to inform our X-ray likelihood function. The prior support of the pulsar radius is truncated at 16 km to ensure coverage of current dense matter models. We assume conservative priors on instrument calibration uncertainty. We constrain the equatorial radius and mass of PSR J0740+6620 to be ${12.39}_{-0.98}^{+1.30}$ km and ${2.072}_{-0.066}^{+0.067}$ M ⊙ respectively, each reported as the posterior credible interval bounded by the 16% and 84% quantiles, conditional on surface hot regions that are non-overlapping spherical caps of fully ionized hydrogen atmosphere with uniform effective temperature; a posteriori, the temperature is ${\mathrm{log}}_{10}(T\,[{\rm{K}}])={5.99}_{-0.06}^{+0.05}$ for each hot region. All software for the X-ray modeling framework is open-source and all data, model, and sample information is publicly available, including analysis notebooks and model modules in the Python language. Furthermore, our marginal likelihood function of mass and equatorial radius is proportional to the marginal joint posterior density of those parameters (within the prior support) and can thus be computed from the posterior samples.

79 ASTRONOMY AND ASTROPHYSICS↗
NASA NTRS2016

Marginalizing Instrument Systematics in HST WFC3 Transit Light Curves

Hubble Space Telescope (HST) Wide Field Camera 3 (WFC3) infrared observations at 1.1-1.7 microns probe primarily the H2O absorption band at 1.4 microns, and have provided low-resolution transmission spectra for a wide range of exoplanets. We present the application of marginalization based on Gibson to analyze exoplanet transit light curves obtained from HST WFC3 to better determine important transit parameters such as "ramp" probability (R (sub p)) divided by "ramp" total (R (sub asterisk)), which are important for accurate detections of H2O. We approximate the evidence, often referred to as the marginal likelihood, for a grid of systematic models using the Akaike Information Criterion. We then calculate the evidence-based weight assigned to each systematic model and use the information from all tested models to calculate the final marginalized transit parameters for both the band-integrated and spectroscopic light curves to construct the transmission spectrum. We find that a majority of the highest weight models contain a correction for a linear trend in time as well as corrections related to HST orbital phase. We additionally test the dependence on the shift in spectral wavelength position over the course of the observations and find that spectroscopic wavelength shifts delta (sub lambda) times lambda) best describe the associated systematic in the spectroscopic light curves for most targets while fast scan rate observations of bright targets require an additional level of processing to produce a robust transmission spectrum. The use of marginalization allows for transparent interpretation and understanding of the instrument and the impact of each systematic evaluated statistically for each data set, expanding the ability to make true and comprehensive comparisons between exoplanet atmospheres.

Wakeford, H. R.↗
NASA NTRS1994

On determining the spectrum of primordial inhomogeneity from the COBE DMR sky maps: Results of two-year data analysis

A new technique of Fourier analysis on a cut sky has been applied to the two-year Cosmic Background Explorer (COBE) Differential Microwave Radiometer (DMR) 53 and 90 GHz sky maps. The Bayesian power spectrum estimation results are consistent with the Harrison-Zel'dovich n = 1 model. The maximum likelihood estimates of the usual parameters defining the power spectrum of primordial perturbations are n = 1.22 (1.02) and Q(sub rms-PS) = 17 (20) microK including (excluding) the quadrupole. A spectral-index-independent normalization is naturally expressed for the two-year maps in terms of the multipole amplitude a(sub 9) = 8.2 (8.3) microK (to approximately 12 sigma significance). The marginal likelihood function on n obtained by intergration with respect to a(sub 9) renders n = 1.17 +/- 0.31 (0.96 +/- 0.36).

Gorski, K. M.↗
NASA NTRS2016

Developing Atmospheric Retrieval Methods for Direct Imaging Spectroscopy of Gas Giants in Reflected Light I: Methane Abundances and Basic Cloud Properties

Reflected light spectroscopy and photometry of cool, directly imaged extrasolar giant planets are expected to be performed in the next decade by space-based telescopes equipped with optical wavelength coronagraphs and integral field spectrographs, such as the Wide-Field Infrared Survey Telescope (WFIRST). We are developing a new atmospheric retrieval methodology to help assess the science return and inform the instrument design for such future missions, and ultimately interpret the resulting observations. Our retrieval technique employs an albedo model coupled with both a Markov chain Monte Carlo Ensemble Sampler (emcee) and a multimodal nested sampling algorithm (MultiNest) to map the posterior distribution. This combination makes the global evidence calculation more robust for any given model, and highlights possible discrepancies in the likelihood maps. Here we apply this methodology to simulated spectra of cool giant planets. As a proof-of-concept, our current atmospheric model contains 1 or 2 cloud layers, methane as a major absorber, and a H2-He background gas. This 6-to-9 parameter model is appropriate for Jupiter-like planets and can be easily expanded in the future. In addition to deriving the marginal likelihood distribution and confidence intervals for the model parameters, we perform model selection to determine the significance of methane and cloud detection as a function of expected signal-to-noise, in the presence of spectral noise correlations. After internal validation, the method is applied to realistic reflected-light spectra of Jupiter, Saturn, and HD 99492 c, a likely observing target. We find that the presence or absence of clouds and methane can be determined with high accuracy, while parameters uncertainties are model-dependent.

spectroscopy↗
NASA NTRS2022

Comparison of Likelihood Methods for Generalized Linear Mixed Models with Application to Quiet Supersonic Flights 2018 Data

Repeated measurement will be a feature of the survey data collected during the Quesst missionX-59 community response tests (CRT). Since each participant will report his or her categorical level of annoyance in response to multiple events, the responses from any single individual may be correlated with one another. Several models within the class of generalized linear mixed models (GLMM) are pertinent to the analysis of correlated categorical outcomes; the random intercept logistic regression model is one example. Both Bayesian and frequentist methods for fitting these models are available, with frequentist methods relying on some form of approximation (of either an integral or the integrand) that appears in the marginal likelihood function. Given several anticipated similarities of the X-59 CRT data to data collected during a past risk reduction, Quiet Supersonic Flights 2018 (QSF18), this short note is intended to create awareness. It documents an instance in which a reported population average dose-response relationship derived from QSF18 single event data was distorted by the integral approximation applied in likelihood-based methods. We review some of the available literature on the topic, compare the outputs of several different computational approaches implemented in available statistical software, and present simple corrective actions that may be useful during the Quesst mission.

dose-response model↗
DOE OSTI2022

Bayesian operator inference for data-driven reduced-order modeling

This work proposes a Bayesian inference method for the reduced-order modeling of time-dependent systems. Informed by the structure of the governing equations, the task of learning a reduced-order model from data is posed as a Bayesian inverse problem with Gaussian prior and likelihood. The resulting posterior distribution characterizes the operators defining the reduced-order model, hence the predictions subsequently issued by the reduced-order model are endowed with uncertainty. The statistical moments of these predictions are estimated via a Monte Carlo sampling of the posterior distribution. Since the reduced models are fast to solve, this sampling is computationally efficient. Furthermore, the proposed Bayesian framework provides a statistical interpretation of the regularization term that is present in the deterministic operator inference problem, and the empirical Bayes approach of maximum marginal likelihood suggests a selection algorithm for the regularization hyperparameters. The proposed method is demonstrated on two examples: the compressible Euler equations with noise-corrupted observations, and a single-injector combustion process.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Physics-Informed Gaussian Process Regression for States Estimation and Forecasting in Power Grids

Real-time state estimation and forecasting are critical for the efficient operation of power grids. In this paper, a physics-informed Gaussian process regression (PhI-GPR) method is presented and used for forecasting and estimating the phase angle, angular speed, and wind mechanical power of a three-generator power grid system using sparse measurements. In standard data-driven Gaussian process regression (GPR), parameterized models for the prior statistics are fit by maximizing the marginal likelihood of observed data. In the PhI-GPR method, we propose to compute the prior statistics offline by solving stochastic differential equations (SDEs) governing the power grid dynamics. The short-term forecast of a power grid system dominated by wind generation is complicated by the stochastic nature of the wind and the resulting uncertainty in wind mechanical power. Here, we assume that the power grid dynamics are governed by swing equations, with the wind mechanical power fluctuating randomly in time. We solve these equations for the mean and covariances of the power grid states using the Monte Carlo simulation method. We demonstrate that the proposed PhI-GPR method can accurately forecast and estimate observed and unobserved states. For the considered problem, PhI-GPR has computational advantages over the ensemble Kalman filter (EnKF) method: In PhI-GPR, ensembles are computed offline and independently of the data acquisition process, whereas for EnFK, ensembles are computed online with data acquisition, rendering real-time forecast more challenging. We also demonstrate that the PhI-GPR forecast is more accurate than the EnKF forecast when the random mechanical wind power is non-Markovian. In contrast, the two methods produce similar forecasts for the Markovian mechanical wind power. For observed states, we show that PhI-GPR provides a forecast comparable to the standard data-driven GPR; both forecasts are significantly more accurate than the autoregressive integrated moving average (ARIMA) forecast. We also show that the ARIMA forecast is more sensitive to observation frequency and measurement errors than the PhI-GPR forecast.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2020

Farmers’ Perceptions of Marginal Land for Biofuel Crops

The dataset includes the survey results about farmers’ perceptions of marginal land availability and the likelihood of a land pixel being marginal based on a machine learning model trained from the survey. Two spreadsheet files are the farmer and farm characteristics (marginal_land_survey_data_shared.xlsx), and the existing land use of marginal lands (land_use_info_sharing.xlsx). Note: the blank cells in these two spreadsheets mean missing values in the survey response. The GeoTiff file includes two bands, one the marginal land likelihood in the Midwestern states (0-1), the other the dominant reason of land marginality (0-5; 0 for farm size, 1 for growing season precipitation, 2 for root zone soil water capacity, 3 for average slope, 4 for growing season mean temperature, and 5 for growing season diurnal range of temperature). To read the data, please use a GIS software such as ArcGIS or QGIS.

marginal land↗
DOE OSTI2021

Optimal Bayesian supervised domain adaptation for RNA sequencing data

Abstract Motivation When learning to subtype complex disease based on next-generation sequencing data, the amount of available data is often limited. Recent works have tried to leverage data from other domains to design better predictors in the target domain of interest with varying degrees of success. But they are either limited to the cases requiring the outcome label correspondence across domains or cannot leverage the label information at all. Moreover, the existing methods cannot usually benefit from other information available a priori such as gene interaction networks. Results In this article, we develop a generative optimal Bayesian supervised domain adaptation (OBSDA) model that can integrate RNA sequencing (RNA-Seq) data from different domains along with their labels for improving prediction accuracy in the target domain. Our model can be applied in cases where different domains share the same labels or have different ones. OBSDA is based on a hierarchical Bayesian negative binomial model with parameter factorization, for which the optimal predictor can be derived by marginalization of likelihood over the posterior of the parameters. We first provide an efficient Gibbs sampler for parameter inference in OBSDA. Then, we leverage the gene-gene network prior information and construct an informed and flexible variational family to infer the posterior distributions of model parameters. Comprehensive experiments on real-world RNA-Seq data demonstrate the superior performance of OBSDA, in terms of accuracy in identifying cancer subtypes by utilizing data from different domains. Moreover, we show that by taking advantage of the prior network information we can further improve the performance. Availability and implementation The source code for implementations of OBSDA and SI-OBSDA are available at the following link. https://github.com/SHBLK/BSDA. Supplementary information Supplementary data are available at Bioinformatics online.

Biochemistry & Molecular Biology↗