Search NASA⌕ Search

SEARCH · Search NASA

Results for “Exponential weighting”

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 73 records · Page 4
All sourcesNASA NTRSDOE OSTI
DOE OSTI2013

Approximation of Müntz-Szász type in weighted spaces

The paper looks at whether a system of exponentials exp(-λ{sub n}t), Reλ{sub n}>0, is complete in various function spaces on the half-line R{sub +}. Wide classes of Banach spaces E and F of functions on R{sub +} are described such that this system is complete in E and F simultaneously. A test is established to determine when this system is complete in the weighted spaces C{sub 0} and L{sup p} with weight (1+t){sup α} on R{sub +}, for α<0 and α<-1, respectively. Bibliography: 18 titles.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2012

A novel conformity index for intensity modulated radiation therapy plan evaluation

Purpose: Intensity modulated radiation therapy (IMRT) has gained popularity in the treatment of cancers. Manual evaluation of IMRT plans for head-and-neck cancers has been especially challenging necessitating efficient and objective assessment tools. In this work, the authors address this issue by developing a personalized conformity index (CI) for comparison of IMRT plans for head-and-neck cancers and evaluating its plan quality discerning power in comparison with other widely used CIs. Methods: A two-dimensional CI with dose and distance incorporated (CI{sub DD}) was developed using the MATLAB program language, to quantify the planning target volume (PTV) coverage. Valuable information contained in the digital imaging and communication in medicine (DICOM) RT objects were harvested for computation of each of the CI{sub DD} components. Apart from the dose penalty factor, a distance-based exponential function was employed by varying the penalty weight associated with the location of cold spots within the PTV. With the goal of deriving a customized penalty factor, the distances between individual pixel and its nearest PTV boundary was found. Using the exponential function, the impact of distance penalty was substantially larger for cold spots closer to the PTV centroid but petered out quickly wherever they were situated in the vicinity of PTV border. In order to evaluate the CI{sub DD} scoring system, three CT image data sets of nasopharyngeal carcinoma (NPC) patients were collected. Ten IMRT plans with degrading qualities were generated from each dataset and were ranked based on CI{sub DD} and other existing indices. The coefficient of variance was calculated for each dataset to compare the degree of variation. Results: The CI{sub DD} scoring system that considered spatial importance of each voxel within the PTV was successfully developed. The results demonstrated that the CI{sub DD} including four discrete factors could provide accurate rankings of plan quality by examining the relative importance of each cold spot within the PTVs. Apart from the dose penalty factor, a distance-based exponential function was employed taking the specific tumor geometry into account. Compared with other commonly used CIs, the CI{sub DD} resulted in the largest coefficient of variance among the ten IMRT plans for each dataset, indicating that its discerning power was the best among the CIs being compared. Conclusions: The CI{sub DD} scoring system was successfully developed to incorporate patient-specific spatial dose information and provide a geometry-based physical index for comparison of IMRT plans for head-and-neck cancers. By taking individual tumor geometry into account, the superiority of CI{sub DD} in plan discerning power was demonstrated. The use of CI{sub DD} could provide an effective means of benchmarking performance, reducing treatment plan variability, and advancing the quality of current IMRT planning.

60 APPLIED LIFE SCIENCES↗
NASA NTRS1983

An improvement in SAR image interpretability provided by post-correlation signal processing

A presentation on the basis of subjective analysis of computer-generated SAR imagery depicts the improvement in interpretability obtained by post-correlation signal processing. A parametric study was conducted to determine the improvement in interpretability obtained by the application of signal weighting functions on the post-processed returns. The results suggest that a marked improvement in interpretability results from symmetrizing the exponential distribution of the fading signal. Preliminary analysis indicates that signal weighting improves the contrast ratio between the mean value of adjacent homogeneous regions in a SAR scene.

Matthews, N. D.↗
DOE OSTI2020

Mapped Moments to a Cartesian Grid (MMCG) Value-Added Product Report

Objective analysis (OA) is a method of mapping unstructured data to a structured grid. In the context of scanning radar data, OA is used to interpolate data in antenna coordinates (range, azimuth, and elevation) onto a regularly spaced Cartesian grid (Trapp and Doswell 2000). The U.S. Department of Energy Atmospheric Radiation Measurement (ARM) user facility’s Mapped Moments to a Cartesian Grid (MMCG) Value-Added Product (VAP) uses the Python ARM Radar Toolkit (Py-ART), a data model-driven interactive architecture for working with weather radar data, to map the data to a Cartesian grid (Helmus and Collis 2016). MMCG, with Py-ART built in, has the ability to take radar data in antenna coordinates and map the gates to a Cartesian grid using inverse distance weight functions such as Cressman (square) and Barnes (exponential), but also can filter the data during the interpolation. MMCG also allows arbitrary formulations for the radius of influence, which are matched to particular radar scanning strategies. This creates a complex parameter space for optimizing the retention of storm structure detail while minimizing artifacts. MMCG takes data processed with ARM’s Corrected Precipitation Radar Moments in Antenna Coordinates (CMAC) VAP and maps it to a Cartesian grid as the output product. A variety of fields that have been mapped to the Cartesian grid are then saved to plots to complement each grid file.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2025

Triangle Method for Dense ReLU Layers [SWR-25-72]

This software is an implementation of the methods for initializing and training neural networks to be more efficient per parameter, described more fully below and in the related publication: In theory, depth should make a ReLU network EXPONENTIALLY more efficient by enabling it to produce an exponential number of piecewise linear sections in its output. This reasoning is largely based on the work of mathematicians that have hand-constructed networks that make good use of depth. In practice however, even very deep ReLU networks that have been randomly initialized will behave identically to their shallow counterparts - missing an entire exponential dimension of efficiency. The triangle method is a first attempt at realizing the exponential potential of deep networks. Instead of randomly setting weights, we force pairs of neurons in each layer learn to build triangles (i.e. functions from [0,1] -> [0,1] that look like triangles). This is a very efficient pattern for generating lots of linear pieces because composing two triangular functions doubles the number of pieces with each composition. The triangle method is more than just a different initialization, it is a new paradigm of training. Instead of making direct updates to the matrix weights, we do an extra step of backpropagation to collect the derivatives of the loss function with respect to the shapes of the triangles, training them to tilt left or right. This process essentially holds the networks hand throughout the loss landscape and forces it to always use depth effectively by producing triangular shapes internally. This can produce several orders of magnitude of improvement on convex one-dimensional regression problems. Much more theoretical work is needed to realize its full potential beyond this context, but the implementation in this repository will still work in arbitrary numbers of dimensions. The file Triangle_Method.py is a generalized form of the method that will build each neuron its own custom 1-d convex activation function (with exponential efficiency). Example usage on one dimensional problems can be found in Example_Usage.ipynb and an example of using this in a real neural network can be found in Example_VGG16_CIFAR10.ipynb.

Milkert, Max [National Renewable Energy Laboratory↗
DOE OSTI2014

Chemical kinetics parameters and model validation for the gasification of PCEA nuclear graphite

A series of gasification experiments, using two right cylinder specimens (~ 12.7 x 25.4 mm and 25.4 x 25.4 mm) of PCEA nuclear graphite in ambient airflow, measured the total gasification flux at weight losses up to 41.5% and temperatures (893-1015 K) characteristics of those for in-pores gasification Mode (a) and in-pores diffusion-limited Mode (b). The chemical kinetics parameters for the gasification of PCEA graphite are determined using a multi-parameters optimization algorithm from the measurements of the total gasification rate and transient weight loss in experiments. These parameters are: (i) the pre-exponential rate coefficients and the Gaussian distributions and values of specific activation energies for adsorption of oxygen and desorption of CO gas; (ii) the specific activation energy and pre-exponential rate coefficient for the breakup of stable un-dissociated C(O2) oxygen radicals to form stable (CO) complexes; (iii) the specific activation energy and pre-exponential coefficient for desorption of CO2 gas and; (iv) the initial surface area of reactive free sites per unit mass. This area is consistently 13.5% higher than that for nuclear graphite grades of NBG-25 and IG-110 and decreases inversely proportional with the square root of the initial mass of the graphite specimens in the experiments. Experimental measurements successfully validate the chemical-reactions kinetics model that calculates continuous Arrhenius curves of the total gasification flux and the production rates of CO and CO2 gases. The model results at different total weight losses agree well with measurements and expand beyond the temperatures in the experiments to the diffusion-limited mode of gasification. Also calculated are the production rates of CO and CO2 gases and their relative contributions to the total gasification rate in the experiments as functions of temperature, for total weight losses of 5% and 10%.

kinetics↗
NASA NTRS1972

The DOPEX Code: an Application of the Method of Steepest Descent to Laminated-Shield-Weight Optimization with Several Constraints

A two- or three-constraint, two-dimensional radiation shield weight optimization procedure and a computer program, DOPEX, is described. The DOPEX code uses the steepest descent method to alter a set of initial (input) thicknesses for a shield configuration to achieve a minimum weight while simultaneously satisfying dose constaints. The code assumes an exponential dose-shield thickness relation with parameters specified by the user. The code also assumes that dose rates in each principal direction are dependent only on thicknesses in that direction. Code input instructions, FORTRAN 4 listing, and a sample problem are given. Typical computer time required to optimize a seven-layer shield is about 0.1 minute on an IBM 7094-2.

Lahti, G. P.↗
NASA NTRSDate not supplied

Affordable Development Strategy for NEP Nuclear Systems

One nuclear electric propulsion (NEP) reactor systems under consideration is a hydride moderated thermal spectrum reactor fueled by high assay low enriched uranium (HALEU). While such a reactor is expected to yield the lightest HALEU reactor design, its development challenges grow exponentially with increasing mission demands, most notably power output, specific weight of the overall system (which may require operation at temperatures exceeding 1200 K), service lifetime, and human-rated reliability. Two of the greatest cost drivers are full-powered nuclear demonstrations and extensive material development campaigns, so it is important to consider options that can minimize the need for or complexity of such tasks. This paper discusses a structured framework being developed for assessing how NEP design choices, such as materials selection, neutronic features, and heat-removal technologies, can translate into project risk and how project performance goals can be traded with development cost.Reactors operating at high temperatures often require cutting-edge heat transfer technologies and creep-resistant materials. Use of new materials in high temperature reactors brings additional complication beyond those common to any new space materials development campaign. For example, such materials may not possess necessary neutronic cross-sectional or neutronic irradiation data. Similarly, use of new materials may significantly influence core neutronics; in some extreme cases, neutronic reactivity feed-back of certain new materials can vary during their service life as radiation damage impacts the scattering cross-section. In an affordable development approach, high fidelity modeling and simulation tools are used to identify and characterize potential ‘knees-in-the-curves’ in the relationship that exists between the mission characteristics and the project risk. Of particular significance is use of modern uncertainty management and variance reduction methods to perform gap analyses that feed into phenomena identification and ranking tables (PIRT) commonly used to communicate nuclear readiness levels. Model-based measurements techniques are used to design sub-scale experiments as a substitute to minimize orcompletely eliminate the need for nuclear demonstrations.This paper will describe the approach and present preliminary results. It will lay the groundwork for developing a set of metrics that can be broadly characterized as system nuclear readiness levels and advancement degree of difficulty for nuclear systems. Equally importantly, a goal of this paper is to initiate a dialogue among stakeholders on what is the sufficient level of maturity that is required for launching a demonstration unit.

Dasari V Rao↗
DOE OSTI2025

Feature learning and generalization in deep networks with orthogonal weights

Fully-connected deep neural networks with weights initialized from independent Gaussian distributions can be tuned to criticality, which prevents the exponential growth or decay of signals propagating through the network. However, such networks still exhibit fluctuations that grow linearly with the depth of the network, which may impair the training of networks with width comparable to depth. We show analytically that rectangular networks with tanh activations and weights initialized from the ensemble of orthogonal matrices have corresponding preactivation fluctuations which are independent of depth, to leading order in inverse width. Moreover, we demonstrate numerically that, at initialization, all correlators involving the neural tangent kernel (NTK) and its descendants at leading order in inverse width—which govern the evolution of observables during training—saturate at a depth of ~20, rather than growing without bound as in the case of Gaussian initializations. We speculate that this structure preserves finite-width feature learning while reducing overall noise, thus improving both generalization and training speed in deep networks with depth comparable to width. We provide some experimental justification by relating empirical measurements of the NTK to the superior performance of deep non-linear orthogonal networks trained under full-batch gradient descent on the MNIST and CIFAR-10 classification tasks.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS1993

Vibrational nonequilibrium effects on diatomic dissociation rates

The collision-induced dissociation rate of diatomic molecules from a ladder of rotational and anharmonic vibrational states is developed, and the correction for vibrational nonequilibrium is considered. The result is similar to an analytic correction derived by Hammerling et al. (1959) for harmonic oscillators. An empirical correction algorithm suggested by Park (1987, 1990) gives similar results when vibrational temperature is comparable to kinetic temperature but underestimates the dissociation rate when vibrational temperature is small compared with the kinetic temperature. This algorithm uses an effective temperature in the experimentally determined Arrhenius expression for the rate coefficient, which is a weighted average of the vibrational and kinetic temperature, whereas theory indicates that kinetic temperature should appear only in the exponential term of the Arrhenius expression. Nevertheless, an effective temperature can always be found that will numerically duplicate the proper rate coefficient at any given condition, but a constant weighting factor cannot be expected to provide this. However, the algorithm can he adjusted to give reasonable results over a range of conditions if the geometric weighting factor is taken to be a simple linear function of the ratio of vibrational to kinetic temperature in the gas.

Hansen, C. F.↗
NASA NTRS1993

One-parameter scaling and exponential-sum fitting for water vapor and CO2 infrared transmission functions

A medium-sized band model for water vapor and CO2 absorption is developed using the one-parameter scaling approximation. The infrared spectrum is divided into 10 bands. The Planck-weighted diffuse transmittance is reduced to a function dependent only upon the scaled absorber amount and fit by an exponential sum. By selecting specific sets of absorption coefficients for exponential-sum fitting, computations of fluxes and cooling rate are made very fast. Compared to a broadband model, the accuracy, speed, and versatility are all enhanced. With absorption due to water vapor line, continuum, CO2 as well as O3 included, the parameterization introduces an error of less than 1.5 W/sq m in fluxes and less than 0.15 C/day in the tropospheric and lower stratospheric cooling rates.

Chou, Ming-Dah↗
DOE OSTI2017

Cluster decomposition of full configuration interaction wave functions: A tool for chemical interpretation of systems with strong correlation

Approximate full configuration interaction (FCI) calculations have recently become tractable for systems of unforeseen size, thanks to stochastic and adaptive approximations to the exponentially scaling FCI problem. The result of an FCI calculation is a weighted set of electronic configurations, which can also be expressed in terms of excitations from a reference configuration. The excitation amplitudes contain information on the complexity of the electronic wave function, but this information is contaminated by contributions from disconnected excitations, i.e., those excitations that are just products of independent lower-level excitations. The unwanted contributions can be removed via a cluster decomposition procedure, making it possible to examine the importance of connected excitations in complicated multireference molecules which are outside the reach of conventional algorithms. We present an implementation of the cluster decomposition analysis and apply it to both true FCI wave functions, as well as wave functions generated from the adaptive sampling CI algorithm. Furthermore, the cluster decomposition is useful for interpreting calculations in chemical studies, as a diagnostic for the convergence of various excitation manifolds, as well as as a guidepost for polynomially scaling electronic structure models. Applications are presented for (i) the double dissociation of water, (ii) the carbon dimer, (iii) the π space of polyacenes, and (iv) the chromium dimer. While the cluster amplitudes exhibit rapid decay with an increasing rank for the first three systems, even connected octuple excitations still appear important in Cr 2 , suggesting that spin-restricted single-reference coupled-cluster approaches may not be tractable for some problems in transition metal chemistry.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2017

Using infinite-volume, continuum QED and lattice QCD for the hadronic light-by-light contribution to the muon anomalous magnetic moment

In our previous work, the connected and leading disconnected hadronic light-by-light contributions to the muon anomalous magnetic moment (g — 2) have been computed using lattice QCD ensembles corresponding to physical pion mass generated by the RBC/UKQCD Collaboration. However, the calculation is expected to suffer from a significant finite-volume error that scales like 1/L 2 where L is the spatial size of the lattice. In this paper, we demonstrate that this problem is cured by treating the muon and photons in infinite-volume, continuum QED, resulting in a weighting function that is precomputed and saved with affordable cost and sufficient accuracy. We present numerical results for the case when the quark loop is replaced by a muon loop, finding the expected exponential approach to the infinite volume limit and consistency with the known analytic result. Here, we have implemented an improved weighting function which reduces both discretization and finite-volume effects arising from the hadronic part of the amplitude.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2014

Intercomparison of Large-eddy Simulations of Arctic Mixed-phase Clouds: Importance of Ice Size Distribution Assumptions

Large-eddy simulations of mixed-phase Arctic clouds by 11 different models are analyzed with the goal of improving understanding and model representation of processes controlling the evolution of these clouds. In a case based on observations from the Indirect and Semi-Direct Aerosol Campaign (ISDAC), it is found that ice number concentration, Ni, exerts significant influence on the cloud structure. Increasing Ni leads to a substantial reduction in liquid water path (LWP) and potential cloud dissipation, in agreement with earlier studies. By comparing simulations with the same microphysics coupled to different dynamical cores as well as the same dynamics coupled to different microphysics schemes, it is found that the ice water path (IWP) is mainly controlled by ice microphysics, while the inter-model differences in LWP are largely driven by physics and numerics of the dynamical cores. In contrast to previous intercomparisons, all models here use the same ice particle properties (i.e., mass-size, mass-fall speed, and mass-capacitance relationships) and a common radiation parameterization. The constrained setup exposes the importance of ice particle size distributions (PSD) in influencing cloud evolution. A clear separation in LWP and IWP predicted by models with bin and bulk microphysical treatments is documented and attributed primarily to the assumed shape of ice PSD used in bulk schemes. Compared to the bin schemes that explicitly predict the PSD, schemes assuming exponential ice PSD underestimate ice growth by vapor deposition and overestimate mass-weighted fall speed leading to an underprediction of IWP by a factor of two in the considered case.

Arctic↗
NASA NTRS1973

DOPEX-1D2C: A one-dimensional, two-constraint radiation shield optimization code

A one-dimensional, two-constraint radiation sheild weight optimization procedure and a computer program, DOPEX-1D2C, is described. The DOPEX-1D2C uses the steepest descent method to alter a set of initial (input) thicknesses of a spherical shield configuration to achieve a minimum weight while simultaneously satisfying two dose-rate constraints. The code assumes an exponential dose-shield thickness relation with parameters specified by the user. Code input instruction, a FORTRAN-4 listing, and a sample problem are given. Typical computer time required to optimize a seven-layer shield is less than 1/2 minute on an IBM 7094.

Lahti, G. P.↗
NASA NTRS2014

Intercomparison of Large-Eddy Simulations of Arctic Mixed-Phase Clouds: Importance of Ice Size Distribution Assumptions

Large-eddy simulations of mixed-phase Arctic clouds by 11 different models are analyzed with the goal of improving understanding and model representation of processes controlling the evolution of these clouds. In a case based on observations from the Indirect and Semi-Direct Aerosol Campaign (ISDAC), it is found that ice number concentration, Ni, exerts significant influence on the cloud structure. Increasing Ni leads to a substantial reduction in liquid water path (LWP), in agreement with earlier studies. In contrast to previous intercomparison studies, all models here use the same ice particle properties (i.e., mass-size, mass-fall speed, and mass-capacitance relationships) and a common radiation parameterization. The constrained setup exposes the importance of ice particle size distributions (PSDs) in influencing cloud evolution. A clear separation in LWP and IWP predicted by models with bin and bulk microphysical treatments is documented and attributed primarily to the assumed shape of ice PSD used in bulk schemes. Compared to the bin schemes that explicitly predict the PSD, schemes assuming exponential ice PSD underestimate ice growth by vapor deposition and overestimate mass-weighted fall speed leading to an underprediction of IWP by a factor of two in the considered case. Sensitivity tests indicate LWP and IWP are much closer to the bin model simulations when a modified shape factor which is similar to that predicted by bin model simulation is used in bulk scheme. These results demonstrate the importance of representation of ice PSD in determining the partitioning of liquid and ice and the longevity of mixed-phase clouds.

clouds↗