Search NASA⌕ Search

SEARCH · Search NASA

Results for “maximum likelihood estimation”

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 577 records · Page 32

Stochastic Spectral Descent for Discrete Graphical Models

Interest in deep probabilistic graphical models has in-creased in recent years, due to their state-of-the-art performance on many machine learning applications. Such models are typically trained with the stochastic gradient method, which can take a significant number of iterations to converge. Since the computational cost of gradient estimation is prohibitive even for modestly sized models, training becomes slow and practically usable models are kept small. In this paper we propose a new, largely tuning-free algorithm to address this problem. Our approach derives novel majorization bounds based on the Schatten- norm. Intriguingly, the minimizers of these bounds can be interpreted as gradient methods in a non-Euclidean space. We thus propose using a stochastic gradient method in non-Euclidean space. We both provide simple conditions under which our algorithm is guaranteed to converge, and demonstrate empirically that our algorithm leads to dramatically faster training and improved predictive ability compared to stochastic gradient descent for both directed and undirected graphical models.

97 MATHEMATICS AND COMPUTING↗

Measurement of kernel swelling and buffer densification in irradiated UCO-TRISO particles

Radiation-induced volume changes in the fuel kernels and buffer layers of UCO-TRISO particles irradiated to an average burnup of 16.1% FIMA have been determined. Measurements of particle dimensions were made on polished cross-sections of 56 irradiated particles at several different polish planes. The data were then analyzed to compute the equivalent spherical diameters of the kernels and the various coating layers, and these were compared to the average as-fabricated values to determine changes due to irradiation. Finally, the kernel volume was found to have increased by an average of 26 ± 6%. Buffer volume decreased by an average of 39 ± 2% due to densification.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Parametric Sensitivity and Uncertainty Quantification in the Version 1 of E3SM Atmosphere Model Based on Short Perturbed Parameter Ensemble Simulations

Abstract The atmospheric component of Energy Exascale Earth System Model version 1 has included many new features in the physics parameterizations compared to its predecessors. Potential complex nonlinear interactions among the new features create a significant challenge for understanding the model behaviors and parameter tuning. Using the one‐at‐a‐time method, the benefit of tuning one parameter may offset the benefit of tuning another parameter, or improvement in one target variable may lead to degradation in another target variable. To better understand the Energy Exascale Earth System Model version 1 model behaviors and physics, we conducted a large number of short simulations (three days) in which 18 parameters carefully selected from parameterizations of deep convection, shallow convection, and cloud macrophysics and microphysics were perturbed simultaneously using the Latin hypercube sampling method. From the perturbed parameter ensemble simulations and use of different skill score functions, we identified the most sensitive parameters, quantified how the model responds to changes of the parameters for both global mean and spatial distribution, and estimated the maximum likelihood of model parameter space for a number of important fidelity metrics. Comparison of the parametric sensitivity using simulations of two different lengths suggests that perturbed parameter ensemble using short simulations has some bearing on understanding parametric sensitivity of longer simulations. Results from this analysis provide a more comprehensive picture of the Energy Exascale Earth System Model version 1 behavior. The difficulty in reducing biases in multiple variables simultaneously highlights the need of characterizing model structural uncertainty (so‐called embedded errors) to inform future development efforts.

54 ENVIRONMENTAL SCIENCES↗

Property-Composition-Temperature Modeling of Waste Glass Melt Data Subject to a Randomization Restriction

Properties such as viscosity and electrical conductivity of glass melts are functions of melt temperature as well as glass composition. When measuring such a property for several glasses, the property is typically measured at several temperatures for one glass, then at several temperatures for the next glass, and so on. This data-collection process involves a restriction on randomization, which is referred to as split-plot experiment. The split-plot data structure must be accounted for in developing property-composition-temperature models and the corresponding uncertainty equations for model predictions. Instead of ordinary least squares (OLS) regression methods, generalized least squares (GLS) regression methods using restricted maximum likelihood (REML) estimation must be used. This article describes the methodology for developing property-composition-temperature models and corresponding prediction uncertainty equations using the GLS/REML regression approach. Viscosity data collected on 197 simulated nuclear waste glasses are used to illustrate the GLS/REML methods for developing a viscosity-composition-temperature model and corresponding equations for model prediction uncertainties. The correct results using GLS/REML regression are compared to the incorrect results obtained using OLS regression.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Validation Metrics for Fixed Effects and Mixed-Effects Calibration

The modern scientific process often involves the development of a predictive computational model. To improve its accuracy, a computational model can be calibrated to a set of experimental data. A variety of validation metrics can be used to quantify this process. Some of these metrics have direct physical interpretations and a history of use, while others, especially those for probabilistic data, are more difficult to interpret. In this work, a variety of validation metrics are used to quantify the accuracy of different calibration methods. Frequentist and Bayesian perspectives are used with both fixed effects and mixed-effects statistical models. Through a quantitative comparison of the resulting distributions, the most accurate calibration method can be selected. Two examples are included which compare the results of various validation metrics for different calibration methods. It is quantitatively shown that, in the presence of significant laboratory biases, a fixed effects calibration is significantly less accurate than a mixed-effects calibration. This is because the mixed-effects statistical model better characterizes the underlying parameter distributions than the fixed effects model. The results suggest that validation metrics can be used to select the most accurate calibration model for a particular empirical model with corresponding experimental data.

97 MATHEMATICS AND COMPUTING↗

An overview of component qualification using Bayesian statistics and energy methods.

The below overview is designed to give the reader a limited understanding of Bayesian and Maximum Likelihood (MLE) estimation; a basic understanding of some of the mathematical tools to evaluate the quality of an estimation; an introduction to energy methods and a limited discussion of damage potential. This discussion then goes on to presented a limited presentation as to how energy methods and Bayesian estimation are used together to qualify components. Example problems with solutions have been supplied as a learning aid. Bold letters are used to represent random variables. Un-bolded letter represent deterministic values. A concluding section presents a discussion of attributes and concerns.

99 GENERAL AND MISCELLANEOUS↗

Maximum likelihood synchronizer for binary overlapping PCM/NRZ signals.

A maximum likelihood parameter estimation technique for the self bit synchronization problem is investigated. The input sequence to the bit synchronizer is a sequence of binary overlapping PCM/NRZ signal in the presence of white Gaussian noise with zero mean and known variance. The resulting synchronizer consists of matched filters, a transition device and a weighting function. Finally, the performance is examined by Monte Carlo simulations.

Wang, C. D.↗

The sensitivity of a very long baseline interferometer

The theoretical sensitivity of various methods of acquiring and processing interferometer data are compared. It is shown that for a fixed digital recording capacity one bit quantization of single sideband data filtered with a rectangular bandpass and sampled at the Nyquist rate yields the optimum signal to noise ratio. The losses which result from imperfect bandpass, poor image rejection, approximate methods of fringe rotation, fractional bit correction, and loss of quadrature are discussed. Also discussed is the use of the complex delay function as a maximum likelihood fringe estimator.

Rogers, A. E. E.↗

Discrimination Relative to Measures of Non-Normality

The robustness of discriminant functions to nonnormality is investigated. The performance of procedures relative to measures of the difference between the actual distribution of the observations and the usual assumption of normal densities is assessed. For example, the two population, mixed distributions problem with equal costs of misclassification are considered. The parameters are estimated by maximum likelihood and recently proposed robust methods.

Smith, W. B.↗

On-orbit Systems Identification of Flexible Spacecraft Technol., 1984, Pt. 2 p 465-481

Future spacecraft include configurations which are too flexible to be adequately tested prior to flight and which will require on-orbit systems identification to ensure safe operation of the flight control system. The structural dynamics model will evolve and its accuracy will improve in stages as ground tests of full-scale components and replica-scale models are performed. State Space Modeling and Conditional Maximum Likelihood Parameter Estimation methodology can provide the formal probability-based framework for the process of upgrading a model as additional test results are obtained. Although the number of unknown parameters can be reduced by the use of canonical forms for the stability matrix, the number of unknown model parameters quickly becomes unmanageable unless advantage is taken of the relationship of a much fewer number of global model parameters. Distributed parameter systems or partial differential equation models are one way to take advantage of such global parameters to reduce the number of unknown model parameters.

Taylor, L.↗

Measurement of interaction cross section of cosmic ray Fe nuclei (E4 GeV/N) with Al target

Although the study of relatvistic heavy ion interaction encumbered by the low intensities of the incident heavy nuclei in cosmic ray experiments, however, one can use a suitable structure of detector to attain a result with a reasonable accuracy. A cross section value of Fe(E 4 GeV/N) with Al is presented. CR-39 was used as detector and Al as target. The cross section value is estimated using maximum likelihood method.

Ren, G. X.↗

On-orbit systems identification of flexible spacecraft

The process of on-orbit systems identification of flexible spacecraft is examined in terms of the difficulties that are expected because of the potentially unmanageable number of unknown model parameters due to the very high order system models involved. A Jordon block canonical form and global model parameters are used to reduce the number of unknown parameters to manageable numbers. A Bayesian approach is discussed which enables the merging of theoretical models with ground or on-orbit test results by using Unconditional Maximum Likelihood Parameter Estimation. A Modified Newton-Raphson technique is proposed for determining the model parameter estimates and an expected fit error criterion is recommended for the determination of the model structure and order reduction.

Taylor, L. W., Jr.↗

Multiple symbol differential detection of uncoded and trellis coded MPSK

A differential detection for MPSK, which uses a multiple symbol observation interval, is presented and its performance analyzed and simulated. The technique makes use of maximum-likelihood sequence estimation of the transmitted phases rather than symbol-by-symbol detection as in conventional differential detection. As such the performance of this multiple symbol detection scheme fills the gap between conventional (two-symbol observation) differentially coherent detection of MPSK and ideal coherent of MPSK with differential encoding. The amount of improvement gained over conventional differential detection depends on the number of phases, M, and the number of additional symbol intervals added to the observation. What is particularly interesting is that substantial performance improvement can be obtained for only one or two additional symbol intervals of observation. The analysis and simulation results presented are for uncoded and trellis coded MPSK.

Divsalar, Dariush↗

Multiple-symbol differential detection of MPSK

A differential detection technique for MPSK (multiple-phase shift keying), which uses a multiple-symbol observation interval, is presented, and its performance is analyzed and simulated. The technique makes use of maximum-likelihood sequence estimation of the transmitted phases rather than symbol-by-symbol detection as in conventional differential detection. Thus, the performance of this multiple-symbol detection scheme fills the gap between conventional (two-symbol observation) differentially coherent detection of MPSK and ideal coherent of MPSK with differential encoding. The amount of improvement gained over conventional differential detection depends on the number of phases M and the number of additional symbol intervals added to the observation. What is particularly interesting is that substantial performance improvement can be obtained for only one or two additional symbol intervals of observation. The analysis and simulation results presented are for uncoded MPSK.

Divsalar, Dariush↗

Maximum likelihood tuning of a vehicle motion filter

This paper describes the use of maximum likelihood parameter estimation unknown parameters appearing in a nonlinear vehicle motion filter. The filter uses the kinematic equations of motion of a rigid body in motion over a spherical earth. The nine states of the filter represent vehicle velocity, attitude, and position. The inputs to the filter are three components of translational acceleration and three components of angular rate. Measurements used to update states include air data, altitude, position, and attitude. Expressions are derived for the elements of filter matrices needed to use air data in a body-fixed frame with filter states expressed in a geographic frame. An expression for the likelihood functions of the data is given, along with accurate approximations for the function's gradient and Hessian with respect to unknown parameters. These are used by a numerical quasi-Newton algorithm for maximizing the likelihood function of the data in order to estimate the unknown parameters. The parameter estimation algorithm is useful for processing data from aircraft flight tests or for tuning inertial navigation systems.

Trankle, Thomas L.↗

Multiple-Symbol, Partially Coherent Detection of MPSK

Proposed method of reception of multiple-phase-shift-keyed (MPSK) radio signals involves multiple-symbol, partially coherent detection. Instead of attempting to determine phase of transmitted signal during each symbol period as in coherent detection, receiver acquires signal data during multiple-symbol observation interval, then produces maximum-likelihood-sequence estimate of phases transmitted during interval. Combination of coherent-reception and incoherent-reception decision rules are used.

Simon, Marvin K.↗

Pseudo-coherent demodulation for mobile satellite systems

This paper proposes three so-called pseudo-coherent demodulation schemes for use in land mobile satellite channels. The schemes are derived based on maximum likelihood (ML) estimation and detection of an N-symbol observation of the received signal. Simulation results for all three demodulators are presented to allow comparison with the performance of differential PSK (DPSK) and ideal coherent demodulation for various system parameter sets of practical interest.

Divsalar, Dariush↗