Search NASA⌕ Search

SEARCH · Search NASA

Results for “Likelihood Functions”

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 163 records · Page 9
All sourcesNASA NTRSDOE OSTI
NASA NTRSDate not supplied

Frequency Domain Quasi Maximum Likelihood Identification of Low Order Aeroservoelastic Models from Flight-Test Data

In this paper a quasi maximum likelihood method for estimating a low order model of a flexible vehicle has been developed and demonstrated. The quasi maximum likelihood method uses a large number of sensors to estimate the parameters and covariance in a method consistent with the maximum likelihood filter-error method. The cost function has been defined in a frequency domain to improve computational efficiency and to enable estimation of an unstable plant. The method has been demonstrated using flight data from the National Aeronautics and Space Administration X-56A Multi-Utility Technology Testbed flex wing tests, which include unstable flutter modes. The method was able to effectively estimate the frequency and damping of the dynamics of the aircraft and generate a transfer function that is useful for control system evaluations.

Jeffrey Ouellette↗
NASA NTRS2022

Deep Interacting Multiple Model Filtering

In this paper, a deep learning-based multiple model estimation framework is presented for the state estimation of hybrid dynamical systems from high dimensional observations such as camera images. A low dimensional vector which represents the measurement of the latent dynamical system and its corresponding variance are learned using a deep encoder neural network. An Interacting Multiple Model (IMM) filter is used to generate the latent state estimates and covariances using multiple dynamical models, which can be learned using backpropagation through time. The state estimates of the dynamical system and the corresponding covariance matrix are generated from the latent state estimates and covariance using a deep decoder neural network. The whole network is trained in an end-to-end manner using a loss function which minimizes the negative log-likelihood of the neural network parameters. Simulation results are presented using a 2D bouncing ball example and estimation error statistics are computed which demonstrates the accuracy and consistency of the estimation.

Ghananeel Rotithor↗
NASA NTRS1996

Two-stage model of radon-induced malignant lung tumors in rats: effects of cell killing

A two-stage stochastic model of carcinogenesis is used to analyze lung tumor incidence in 3750 rats exposed to varying regimens of radon carried on a constant-concentration uranium ore dust aerosol. New to this analysis is the parameterization of the model such that cell killing by the alpha particles could be included. The model contains parameters characterizing the rate of the first mutation, the net proliferation rate of initiated cells, the ratio of the rates of cell loss (cell killing plus differentiation) and cell division, and the lag time between the appearance of the first malignant cell and the tumor. Data analysis was by standard maximum likelihood estimation techniques. Results indicate that the rate of the first mutation is dependent on radon and consistent with in vitro rates measured experimentally, and that the rate of the second mutation is not dependent on radon. An initial sharp rise in the net proliferation rate of initiated cell was found with increasing exposure rate (denoted model I), which leads to an unrealistically high cell-killing coefficient. A second model (model II) was studied, in which the initial rise was attributed to promotion via a step function, implying that it is due not to radon but to the uranium ore dust. This model resulted in values for the cell-killing coefficient consistent with those found for in vitro cells. An "inverse dose-rate" effect is seen, i.e. an increase in the lifetime probability of tumor with a decrease in exposure rate. This is attributed in large part to promotion of intermediate lesions. Since model II is preferable on biological grounds (it yields a plausible cell-killing coefficient), such as uranium ore dust. This analysis presents evidence that a two-stage model describes the data adequately and generates hypotheses regarding the mechanism of radon-induced carcinogenesis.

NASA Program Radiation Health↗
NASA NTRS1981

Incorporating spatial context into statistical classification of multidimensional image data

Compound decision theory is employed to develop a general statistical model for classifying image data using spatial context. The classification algorithm developed from this model exploits the tendency of certain ground-cover classes to occur more frequently in some spatial contexts than in others. A key input to this contextural classifier is a quantitative characterization of this tendency: the context function. Several methods for estimating the context function are explored, and two complementary methods are recommended. The contextural classifier is shown to produce substantial improvements in classification accuracy compared to the accuracy produced by a non-contextural uniform-priors maximum likelihood classifier when these methods of estimating the context function are used. An approximate algorithm, which cuts computational requirements by over one-half, is presented. The search for an optimal implementation is furthered by an exploration of the relative merits of using spectral classes or information classes for classification and/or context function estimation.

Bauer, M. E.↗
DOE OSTI2024

Neural simulation-based inference of the neutron star equation of state directly from telescope spectra

Neutron stars provide a unique opportunity to study strongly interacting matter under extreme density conditions. The intricacies of matter inside neutron stars and their equation of state are not directly visible, but determine bulk properties, such as mass and radius, which affect the star's thermal X-ray emissions. However, the telescope spectra of these emissions are also affected by the stellar distance, hydrogen column, and effective surface temperature, which are not always well-constrained. Uncertainties on these nuisance parameters must be accounted for when making a robust estimation of the equation of state. In this study, we develop a novel methodology that, for the first time, can infer the full posterior distribution of both the equation of state and nuisance parameters directly from telescope observations. This method relies on the use of neural likelihood estimation, in which normalizing flows use samples of simulated telescope data to learn the likelihood of the neutron star spectra as a function of these parameters, coupled with Hamiltonian Monte Carlo methods to efficiently sample from the corresponding posterior distribution. Our approach surpasses the accuracy of previous methods, improves the interpretability of the results by providing access to the full posterior distribution, and naturally scales to a growing number of neutron star observations expected in the coming years.

79 ASTRONOMY AND ASTROPHYSICS↗
NASA NTRS1971

A system debugging model.

Consideration of the nature of the 'debugging' process applied to a new complex system during the initial period of its life. During this period failures and errors are corrected as they occur, with resulting improvement in the subsequent performance of the system. One mathematical idealization of this process leads to the assumption that system failure rate is decreasing with time. In practice, the debugging phase is considered completed when the failure rate reaches an equilibrium or constant value. Models are formulated for this phenomenon. Maximum likelihood estimates are obtained for relevant failure rate functions and for the end of the debugging period. A conservative upper confidence bound on the stable failure rate is obtained.

Barlow, R. E.↗
NASA NTRS1975

A random search algorithm for laboratory computers

The small laboratory computer is ideal for experimental control and data acquisition. Postexperimental data processing is often performed on large computers because of the availability of sophisticated programs, but costs and data compatibility are negative factors. Parameter optimization can be accomplished on the small computer, offering ease of programming, data compatibility, and low cost. A previously proposed random-search algorithm ('random creep') was found to be very slow in convergence. A method is proposed (the 'random leap' algorithm) which starts in a global search mode and automatically adjusts step size to speed convergence. A FORTRAN executive program for the random-leap algorithm is presented which calls a user-supplied function subroutine. An example of a function subroutine is given which calculates maximum-likelihood estimates of receiver operating-characteristic parameters from binary response data. Other applications in parameter estimation, generalized least squares, and matrix inversion are discussed.

Curry, R. E.↗
NASA NTRS1979

Estimating crop proportions from remotely sensed data

The classification/pixel-count method for estimating the proportion of wheat in each segment is theoretically biased even if all distributional assumptions are met. Alternative ways to estimate crop proportions are examined and their performance testing is considered. Topics covered include general linear functional estimates, the method of moments, and maximum likelihood estimators.

Feiveson, A. H.↗
NASA NTRS1980

CCD data processor for maximum likelihood feature classification

The paper describes an advanced technology development which utilizes a high speed analog/binary CCD correlator to perform the matrix multiplications necessary to implement onboard feature classification. The matrix manipulation module uses the maximum likelihood classification algorithm assuming a Gaussian probability density function. The module will process 16 element multispectral vectors at rates in excess of 500 thousand multispectral vector elements per second. System design considerations for the optimum use of this module are discussed, test results from initial device fabrication runs are presented, and the performance in typical processing applications is described

Benz, H. F.↗
NASA NTRS1980

The luminosity functions of the 1969 Perseid and Orionid meteor showers

Observations of the 1969 Perseid and Orionid meteor showers are presented and used to derive luminosity functions for the 288 Perseids and 56 Orionids detected. Visual counts were performed under very good to excellent seeing conditions at the times of peak activities, and the brightnesses of the meteors were estimated to the nearest magnitude by comparison with the magnitudes of known objects. Maximum likelihood estimates of the power law index of the luminosity function of 1.56 + or - 0.06 for the Perseids and of 1.85 + or - 0.1 for the Orionids are obtained which are lower than the values found by other investigators. Under the assumption that the luminosity of visual meteors is proportional to their mass, the luminosity function power law may also be used to characterize the mass function.

Krisciunas, K.↗
NASA NTRS2015

Predicting the Redshift 2 H-Alpha Luminosity Function Using [OIII] Emission Line Galaxies

Upcoming space-based surveys such as Euclid and WFIRST-AFTA plan to measure Baryonic Acoustic Oscillations (BAOs) in order to study dark energy. These surveys will use IR slitless grism spectroscopy to measure redshifts of a large number of galaxies over a significant redshift range. In this paper, we use the WFC3 Infrared Spectroscopic Parallel Survey (WISP) to estimate the expected number of H-alpha emitters observable by these future surveys. WISP is an ongoing Hubble Space Telescope slitless spectroscopic survey, covering the 0.8 - 1.65 micrometers wavelength range and allowing the detection of H-alpha emitters up to z approximately equal to 1.5 and [OIII] emitters to z approximately equal to 2.3. We derive the H-alpha-[OIII] bivariate line luminosity function for WISP galaxies at z approximately equal to 1 using a maximum likelihood estimator that properly accounts for uncertainties in line luminosity measurement, and demonstrate how it can be used to derive the H-alpha luminosity function from exclusively fitting [OIII] data. Using the z approximately equal to 2 [OIII] line luminosity function, and assuming that the relation between H-alpha and [OIII] luminosity does not change significantly over the redshift range, we predict the H-alpha number counts at z approximately equal to 2 - the upper end of the redshift range of interest for the future surveys. For the redshift range 0.7 less than z less than 2, we expect approximately 3000 galaxies per sq deg for a flux limit of 3 x 10(exp −16) ergs per sec per sq cm (the proposed depth of Euclid galaxy redshift survey) and approximately 20,000 galaxies per sq deg for a flux limit of approximately 10(exp −16) ergs per sec per sq cm (the baseline depth of WFIRST galaxy redshift survey).

Redshift↗
NASA NTRSDate not supplied

Structured Covariance Gaussian Networks for Orion Crew Module Aerodynamic Uncertainty Quantification

In this paper we propose a new approach for nonlinear regression and uncertainty quantification. The method is based on a pair of neural networks which parameterize mean and dense covariance functions of a multivariate Gaussian process, trained together to maximize the log-likelihood of observing the given data. The covariance matrix is made positive definite at every input by construction. We also propose a sampling approach that produces viable surrogate function realizations from the Gaussian process. We call the proposed model a Structured Covariance Gaussian Network (SCGN). We illustrate the use of SCGNs for learning an aerodynamic response surface with built-in uncertainty for the Orion crew module. We find that SCGN provides an efficient and systematic way to learn nonlinear functional relationships and dense covariances. We compare results to a baseline Gaussian process regressor and observe that the SCGN provides comparable uncertainty descriptions with improved scalability to dataset size. The sample functions generated by SCGN are fast to evaluate online and are therefore convenient for use in trajectory simulations. These results suggest that SCGN may be a viable computational method for aerodynamic uncertainty quantification.

machine learning↗
NASA NTRSDate not supplied

Structured Covariance Gaussian Networks for Orion Crew Module Aerodynamic Uncertainty Quantification

In this paper we propose a new approach for nonlinear regression and uncertainty quantification. The method is based on a pair of neural networks which parameterize mean and dense covariance functions of a multivariate Gaussian process, trained together to maximize the log-likelihood of observing the given data. The covariance matrix is made positive definite at every input by construction. We also propose a sampling approach that produces viable surrogate function realizations from the Gaussian process. We call the proposed model a Structured Covariance Gaussian Network (SCGN). We illustrate the use of SCGNs for learning an aerodynamic response surface with built-in uncertainty for the Orion crew module. We find that SCGN provides an efficient and systematic way to learn nonlinear functional relationships and dense covariances. We compare results to a baseline Gaussian process regressor and observe that the SCGN provides comparable uncertainty descriptions with improved scalability to dataset size. The sample functions generated by SCGN are fast to evaluate online and are therefore convenient for use in trajectory simulations. These results suggest that SCGN may be a viable computational method for aerodynamic uncertainty quantification.

machine learning↗
NASA NTRS1999

Distant Massive Clusters and Cosmology

We present a status report of our X-ray study and analysis of a complete sample of distant (z=0.5-0.8), X-ray luminous clusters of galaxies. We have obtained ASCA and ROSAT observations of the five brightest Extended Medium Sensitivity (EMSS) clusters with z > 0.5. We have constructed an observed temperature function for these clusters, and measured iron abundances for all of these clusters. We have developed an analytic expression for the behavior of the mass-temperature relation in a low-density universe. We use this mass-temperature relation together with a Press-Schechter-based model to derive the expected temperature function for different values of Omega-M. We combine this analysis with the observed temperature functions at redshifts from 0 - 0.8 to derive maximum likelihood estimates for the value of Omega-M. We report preliminary results of this analysis.

Donahue, Megan↗
NASA NTRS1981

Characterization, parameter estimation, and aircraft response statistics of atmospheric turbulence

A nonGaussian three component model of atmospheric turbulence is postulated that accounts for readily observable features of turbulence velocity records, their autocorrelation functions, and their spectra. Methods for computing probability density functions and mean exceedance rates of a generic aircraft response variable are developed using nonGaussian turbulence characterizations readily extracted from velocity recordings. A maximum likelihood method is developed for optimal estimation of the integral scale and intensity of records possessing von Karman transverse of longitudinal spectra. Formulas for the variances of such parameter estimates are developed. The maximum likelihood and least-square approaches are combined to yield a method for estimating the autocorrelation function parameters of a two component model for turbulence.

Mark, W. D.↗
NASA NTRS1979

Maximum likelihood estimation for life distributions with competing failure modes

Systems which are placed on test at time zero, function for a period and die at some random time were studied. Failure may be due to one of several causes or modes. The parameters of the life distribution may depend upon the levels of various stress variables the item is subject to. Maximum likelihood estimation methods are discussed. Specific methods are reported for the smallest extreme-value distributions of life. Monte-Carlo results indicate the methods to be promising. Under appropriate conditions, the location parameters are nearly unbiased, the scale parameter is slight biased, and the asymptotic covariances are rapidly approached.

Sidik, S. M.↗
DOE OSTI2025

Massive νs through the CNN lens: interpreting the field-level neutrino mass information in weak lensing

Modern cosmological surveys probe the Universe deep into the nonlinear regime, where massive neutrinos suppress cosmic structure. Traditional cosmological analyses, which use the 2-point correlation function to extract information, are no longer optimal in the nonlinear regime, and there is thus much interest in extracting beyond-2-point information to improve constraints on neutrino mass. Quantifying and interpreting the beyond-2-point information is thus a pressing task. We study the field-level information in weak lensing convergence maps using convolution neural networks. We find that the network performance increases as higher source redshifts and smaller scales are considered — investigating up to a source redshift of 2.5 and ℓ max ≃ 10 4 — verifying that massive neutrinos leave a distinct effect on weak lensing. However, the performance of the network significantly drops after scaling out the 2-point information from the maps, implying that most of the field-level information can be found in the 2-point correlation function alone. We quantify these findings in terms of the likelihood ratio and also use Integrated Gradient saliency maps to interpret which parts of the map the network is learning the most from. We find that, in the absence of noise, the network extracts a similar amount of information from the most overdense and underdense regions. However, upon adding noise, the information in underdense regions is distorted as noise disproportionately washes out void-like structures.

Golshan, Malika [University of California, Berkele↗