SEARCH · Search NASA

Results for “Function approximation”

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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,063 records · Page 59

All sources NASA NTRS DOE OSTI

DOE OSTI2026

Augmenting subspace optimization methods with linear bandits

In this work, we consider the framework of methods for unconstrained minimization that are, in each iteration, restricted to a model that is only a valid approximation to the objective function on some affine subspace containing an incumbent point. These methods are of practical interest in computational settings where derivative information is either expensive or impossible to obtain. Recent attention has been paid in the literature to employing randomized matrix sketching for generating the affine subspaces within this framework. We consider a relatively straightforward, deterministic augmentation of such a generic subspace optimization method. In particular, we consider a sequential optimization framework where actions consist of one-dimensional linear subspaces and rewards consist of (approximations to) the magnitudes of directional derivatives computed in the direction of the action subspace. Reward maximization in this context is consistent with maximizing lower bounds on descent guaranteed by first-order Taylor models. This sequential optimization problem can be analysed through the lens of dynamic regret. We modify an existing linear upper confidence bound (UCB) bandit method and prove sublinear dynamic regret in the subspace optimization setting. We demonstrate the efficacy of employing this linear UCB method in a setting where forward-mode algorithmic differentiation can provide directional derivatives in arbitrary directions and in a derivative-free setting. For the derivative-free setting, we propose SS-POUNDers, an extension of the derivative-free optimization method POUNDers that employs the linear UCB mechanism to identify promising subspaces. Our numerical experiments suggest a preference, in either computational setting, for employing a linear UCB mechanism within a subspace optimization method.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2023

Hierarchical off-diagonal low-rank approximation of Hessians in inverse problems, with application to ice sheet model initialization

Obtaining lightweight and accurate approximations of discretized objective functional Hessians in inverse problems governed by partial differential equations (PDEs) is essential to make both deterministic and Bayesian statistical large-scale inverse problems computationally tractable. The cubic computational complexity of dense linear algebraic tasks, such as Cholesky factorization, that provide a means to sample Gaussian distributions and determine solutions of Newton linear systems is a computational bottleneck at large-scale. These tasks can be reduced to log-linear complexity by utilizing hierarchical off-diagonal low-rank (HODLR) matrix approximations. In this work, we show that a class of Hessians that arise from inverse problems governed by PDEs are well approximated by the HODLR matrix format. In particular, we study inverse problems governed by PDEs that model the instantaneous viscous flow of ice sheets. In these problems, we seek a spatially distributed basal sliding parameter field such that the flow predicted by the ice sheet model is consistent with ice sheet surface velocity observations. Here, we demonstrate the use of HODLR Hessian approximation to efficiently sample the Laplace approximation of the posterior distribution with covariance further approximated by HODLR matrix compression. Computational studies are performed which illustrate ice sheet problem regimes for which the Gauss–Newton data-misfit Hessian is more efficiently approximated by the HODLR matrix format than the low-rank (LR) format. We then demonstrate that HODLR approximations can be favorable, when compared to global LR approximations, for large-scale problems by studying the data-misfit Hessian associated with inverse problems governed by the first-order Stokes flow model on the Humboldt glacier and Greenland ice sheet.

97 MATHEMATICS AND COMPUTING↗

NASA NTRS1998

Determination of Particle Sources for a Geotail Distribution Function Observed on May 23, 1995

On May 23, 1995, the Comprehensive Plasma Instrumentation (CPI) onboard the Geotail spacecraft observed a complex and structured ion distribution function near the magnetotail midplane at x approximately -10 R(sub E). On the same day, the Wind spacecraft observed a very high density (approximately 40/cubic cm) solar wind and an interplanetary magnetic field (IMF) that was predominantly northward but had several southward turnings. We have inferred the sources of the ions in this distribution function by following approximately 90,000 ion trajectories backward in time using time-dependent electric and magnetic fields obtained from a global MHD (magnetohydrodynamic) simulation. Wind data were used as input for the MHD model. We found that three sources contributed to this distribution: the ionosphere, the plasma mantle which had near-Earth and distant tail components, and the low latitude boundary layer (LLBL). Moreover, distinct structures in the low energy part of the distribution function were found to be associated with individual sources. Structures near 0 deg pitch angle were made up of either ionospheric or plasma mantle ions, while structures near 90 deg pitch angle were dominated by ions from the LLBL source. Particles that underwent nonadiabatic acceleration were numerous in the higher energy part of the ion distribution function, whereas ionospheric and LLBL ions were mostly adiabatic. A large proportion of the near-Earth mantle ions underwent adiabatic acceleration, while most of the distant mantle ions experienced nonadiabatic acceleration.

Ashour-Abdalla, M.↗

NASA NTRS1982

Solution to the backward-Kolmogorov equation for a nonstationary oscillation problem

The transition probability density function of a Markovian approximation of the response amplitude of an oscillator under nonstationary excitation is determined in an analytical form. A solution is presented for the associated backward-Kolmogorov equation by transforming the equation into a form amenable to solution by the method of the separation of variables. This procedure results in a boundary value problem which is then solved by using an infinite series of Laguerre polynomials. It is found that the infinite series solution is equivalent to a closed-form solution involving a Bessel function.

Solomos, G. P.↗

NASA NTRS2008

Communication Limits Due to Photon-Detector Jitter

A theoretical and experimental study was conducted of the limit imposed by photon-detector jitter on the capacity of a pulse-position-modulated optical communication system in which the receiver operates in a photon-counting (weak-signal) regime. Photon-detector jitter is a random delay between impingement of a photon and generation of an electrical pulse by the detector. In the study, jitter statistics were computed from jitter measurements made on several photon detectors. The probability density of jitter was mathematically modeled by use of a weighted sum of Gaussian functions. Parameters of the model were adjusted to fit histograms representing the measured-jitter statistics. Likelihoods of assigning detector-output pulses to correct pulse time slots in the presence of jitter were derived and used to compute channel capacities and corresponding losses due to jitter. It was found that the loss, expressed as the ratio between the signal power needed to achieve a specified capacity in the presence of jitter and that needed to obtain the same capacity in the absence of jitter, is well approximated as a quadratic function of the standard deviation of the jitter in units of pulse-time-slot duration.

Moision, Bruce E.↗

NASA NTRS1989

Objective analysis of pseudostress over the Indian Ocean using a direct-minimization approach

A technique not previously used in objective analysis of meteorological data is used here to produce monthly average surface pseudostress data over the Indian Ocean. An initial guess field is derived and a cost functional is constructed with five terms: approximation to initial guess, approximation to climatology, a smoothness parameter, and two kinematic terms. The functional is minimized using a conjugate-gradient technique, and the weight for the climatology term controls the overall balance of influence between the climatology and the initial guess. Results from various weight combinations are presented for January and July 1984. Quantitative and qualitative comparisons to the subject analysis are made to find which weight combination provides the best results. The weight on the approximation to climatology is found to balance the influence of the original field and climatology.

Legler, David M.↗

NASA NTRS1975

A grid of model atmospheres for metal-deficient giant stars. I

A grid of flux-constant model atmospheres for stars with effective temperatures between 3750 and 6000 K, log g between 0.75 and 3.0, and (A/H) between -3.0 and 0.0 has been constructed. The line absorption is approximated by opacity distribution functions. Metal lines and molecular lines, including those from the infrared bands of CO and CN, are taken into account. The variation of the structure of the models with metal abundance and microturbulence parameter is found to be quite regular and not very drastic. The surface cooling produced by CO is important for all the models with a maximum temperature of 5000 K, while CN mainly causes a backwarming effect but is not very important for solar CNO abundances. The effects of convection, estimated by using the mixing-length approximation, are important only for the coldest models. The models compare very well with models from overlapping regions of other grids. A solar model consistent with the grid models is found to agree satisfactorily with empirical solar model atmospheres. The careful use of scaled solar models for stars with (A/H) approximately equal to zero is justified.

Gustafsson, B.↗

DOE OSTI2020

Confirmation of the PPLB Derivative Discontinuity: Exact Chemical Potential at Finite Temperatures of a Model System

The landmark 1982 paper of Perdew, Parr, Levy, and Balduz (often called PPLB) laid the foundation for our modern understanding of the role of the derivative discontinuity in density functional theory, which drives much development to account for its effects. A simple model for the chemical potential at vanishing temperature played a crucial role in their argument. We investigate the validity of this model in the simplest non-trivial system to which it can be applied and which can be easily solved exactly, the Hubbard dimer. We find exact agreement in the crucial zero-temperature limit, and show the model remains accurate for a significant range of temperatures. We identify how this range depends on the strength of correlations. We extend the model to approximate free energies accounting for the derivative discontinuity, a feature missing in standard semilocal approximations. Here, we provide a correction to this approximation to yield even more accurate free energies. We discuss the relevance of these results for warm dense matter.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2021

Production of light-flavor hadrons in pp collisions at $\sqrt{s}~=~7\text { and }\sqrt{s} = 13 \, \text { TeV} $

The production of $\pi ^{\pm }$ , $\mathrm{K}^{\pm }$ , $\mathrm{K}^{0}_{S}$ , $\mathrm{K}^{*}(892)^{0}$ , $\mathrm{p}$ , $\phi (1020)$ , $\Lambda $ , $\Xi ^{-}$ , $\Omega ^{-}$ , and their antiparticles was measured in inelastic proton–proton (pp) collisions at a center-of-mass energy of $\sqrt{s}$ = 13 TeV at midrapidity ( $|y|<0.5$ ) as a function of transverse momentum ( $p_{\mathrm{T}}$ ) using the ALICE detector at the CERN LHC. Furthermore, the single-particle $p_{\mathrm{T}}$ distributions of $\mathrm{K}^{0}_{S}$ , $\Lambda $ , and $\overline{\Lambda }$ in inelastic pp collisions at $\sqrt{s} = 7$ TeV are reported here for the first time. The $p_{\mathrm{T}}$ distributions are studied at midrapidity within the transverse momentum range $0\le p_{\mathrm{T}}\le 20$ GeV/ c , depending on the particle species. The $p_{\mathrm{T}}$ spectra, integrated yields, and particle yield ratios are discussed as a function of collision energy and compared with measurements at lower $\sqrt{s}$ and with results from various general-purpose QCD-inspired Monte Carlo models. A hardening of the spectra at high $p_{\mathrm{T}}$ with increasing collision energy is observed, which is similar for all particle species under study. The transverse mass and $x_{\mathrm{T}}\equiv 2p_{\mathrm{T}}/\sqrt{s}$ scaling properties of hadron production are also studied. As the collision energy increases from $\sqrt{s}$ = 7–13 TeV, the yields of non- and single-strange hadrons normalized to the pion yields remain approximately constant as a function of $\sqrt{s}$ , while ratios for multi-strange hadrons indicate enhancements. The $p_\mathrm{{T}}$ -differential cross sections of $\pi ^{\pm }$ , $\mathrm {K}^{\pm }$ and $\mathrm {p}$ ( $\overline{\mathrm{p}}$ ) are compared with next-to-leading order perturbative QCD calculations, which are found to overestimate the cross sections for $\pi ^{\pm }$ and $\mathrm{p}$ ( $\overline{\mathrm{p}}$ ) at high $p_\mathrm{{T}}$ .

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

NASA NTRS2000

General Purpose Fortran Program for Discrete-Ordinate-Method Radiative Transfer in Scattering and Emitting Layered Media: An Update of DISORT

This update reports a state-of-the-art discrete ordinate algorithm for monochromatic unpolarized radiative transfer in non-isothermal, vertically inhomogeneous, but horizontally homogeneous media. The physical processes included are Planckian thermal emission, scattering with arbitrary phase function, absorption, and surface bidirectional reflection. The system may be driven by parallel or isotropic diffuse radiation incident at the top boundary, as well as by internal thermal sources and thermal emission from the boundaries. Radiances, fluxes, and mean intensities are returned at user-specified angles and levels. DISORT has enjoyed considerable popularity in the atmospheric science and other communities since its introduction in 1988. Several new DISORT features are described in this update: intensity correction algorithms designed to compensate for the 8-M forward-peak scaling and obtain accurate intensities even in low orders of approximation; a more general surface bidirectional reflection option; and an exponential-linear approximation of the Planck function allowing more accurate solutions in the presence of large temperature gradients. DISORT has been designed to be an exemplar of good scientific software as well as a program of intrinsic utility. An extraordinary effort has been made to make it numerically well-conditioned, error-resistant, and user-friendly, and to take advantage of robust existing software tools. A thorough test suite is provided to verify the program both against published results, and for consistency where there are no published results. This careful attention to software design has been just as important in DISORT's popularity as its powerful algorithmic content.

Tsay, Si-Chee↗

DOE OSTI2021

Measurements of dihadron correlations relative to the event plane in Au+Au collisions at $\sqrt{s_{NN}} =200$ GeV*

Dihadron azimuthal correlations containing a high transverse momentum ($ p_{T} $) trigger particle are sensitive to the properties of the nuclear medium created at RHIC through the strong interactions occurring between the traversing parton and the medium, i.e. jet-quenching. Previous measurements revealed a strong modification to dihadron azimuthal correlations in Au+Au collisions with respect to p+p and d+Au collisions. The modification increases with the collision centrality, suggesting a path-length or energy density dependence to the jet-quenching effect. This paper reports STAR measurements of dihadron azimuthal correlations in mid-central (20%-60%) Au+Au collisions at $ \sqrt{s_{\rm{NN}}} = 200 $ GeV as a function of the trigger particle's azimuthal angle relative to the event plane, $ \phi_{s} = | \phi_{t}- \psi_{{\rm{EP}}}| $. The azimuthal correlation is studied as a function of both the trigger and associated particle $ p_{T} $. The subtractions of the combinatorial background and anisotropic flow, assuming Zero Yield At Minimum (ZYAM), are described. The correlation results are first discussed with subtraction of the even harmonic (elliptic and quadrangular) flow backgrounds. The away-side correlation is strongly modified, and the modification varies with $ \phi_{s} $, with a double-peak structure for out-of-plane trigger particles. The near-side ridge (long range pseudo-rapidity $ \Delta\eta $ correlation) appears to drop with increasing $ \phi_{s} $ while the jet-like component remains approximately constant. The correlation functions are further studied with the subtraction of odd harmonic triangular flow background arising from fluctuations. It is found that the triangular flow, while responsible for the majority of the amplitudes, is not sufficient to explain the $ \phi_{s} $-dependence of the ridge or the away-side double-peak structure. The dropping ridge with $ \phi_{s} $ could be attributed to a $ \phi_{s} $-dependent elliptic anisotropy; however, the physics mechanism of the ridge remains an open question. Even with a $ \phi_{s} $-dependent elliptic flow, the away-side correlation structure is robust. These results, with extensive systematic studies of the dihadron correlations as a function of $ \phi_{s} $, trigger and associated particle $ p_{T} $, and the pseudo-rapidity range $ \Delta\eta $, should provide stringent inputs to help understand the underlying physics mechanisms of jet-medium interactions in high energy nuclear collisions.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

DOE OSTI2023

Linearization errors in discrete goal-oriented error estimation

This paper is concerned with goal-oriented a posteriori error estimation for nonlinear functionals in the context of nonlinear variational problems solved with continuous Galerkin finite element discretizations. A two-level, or discrete, adjoint-based approach for error estimation is considered. The traditional method to derive an error estimate in this context requires linearizing both the nonlinear variational form and the nonlinear functional of interest which introduces linearization errors into the error estimate. In this paper, we investigate these linearization errors. In particular, we develop a novel discrete goal-oriented error estimate that accounts for traditionally neglected nonlinear terms at the expense of greater computational cost. We demonstrate how this error estimate can be used to drive mesh adaptivity. Here, we show that accounting for linearization errors in the error estimate can improve its effectivity for several nonlinear model problems and quantities of interest. We also demonstrate that an adaptive strategy based on the newly proposed estimate can lead to more accurate approximations of the nonlinear functional with fewer degrees of freedom when compared to uniform refinement and traditional adjoint-based approaches.

42 ENGINEERING↗

DOE OSTI2023

Probing reaction channels via reinforcement learning

Abstract Chemical reactions are dynamical processes involving the correlated reorganization of atomic configurations, driving the conversion of an initial reactant into a result product. By virtue of the metastability of both the reactants and products, chemical reactions are rare events, proceeding fleetingly. Reaction pathways can be modelled probabilistically by using the notion of reactive density in the phase space of the molecular system. Such density is related to a function known as the committor function, which describes the likelihood of a configuration evolving to one of the nearby metastable regions. In theory, the committor function can be obtained by solving the backward Kolmogorov equation (BKE), which is a partial differential equation (PDE) defined in the full dimensional phase space. However, using traditional methods to solve this problem is not practical for high dimensional systems. In this work, we propose a reinforcement learning based method to identify important configurations that connect reactant and product states along chemical reaction paths. By shooting multiple trajectories from these configurations, we can generate an ensemble of states that concentrate on the transition path ensemble. This configuration ensemble can be effectively employed in a neural network-based PDE solver to obtain an approximation solution of a restricted BKE, even when the dimension of the problem is very high. The resulting solution provides an approximation for the committor function that encodes mechanistic information for the reaction, paving a new way for understanding of complex chemical reactions and evaluation of reaction rates.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

NASA NTRS1986

Unimolecular diffusion-mediated reactions with a nonrandom time-modulated absorbing barrier

A diffusion-reaction model with time-dependent reactivity is formulated and applied to unimolecular reactions. The model is solved exactly numerically and approximately analytically for the unreacted fraction as a function of time. It is shown that the approximate analytical solution is valid even when the system is far from equilibrium, and when the reactivity probability is more complicated than a square-wave function of time. A discussion is also given of an approach to problems of this type using a stochastically fluctuating reactivity, and the first-passage time for a particular example is derived.

Bashford, D.↗

DOE OSTI2024

Approaching periodic systems in ensemble density functional theory via finite one-dimensional models

Ensemble density functional theory (EDFT) is a generalization of ground-state DFT, which is based on an exact formal theory of finite collections of a system's ground and excited states. EDFT in various forms has been shown to improve the accuracy of calculated energy level differences in isolated model systems, atoms, and molecules, but it is not yet clear how EDFT could be used to calculate band gaps for periodic systems. We extend the application of EDFT toward periodic systems by estimating the thermodynamic limit with increasingly large finite one-dimensional 'particle in a box' systems, which approach the uniform electron gas (UEG). Using ensemble-generalized Hartree and local spin density approximation exchange-correlation functionals, we find that corrections go to zero in the infinite limit, as expected for a metallic system. However, there is a correction to the effective mass, with results comparable to other calculations on 1D, 2D, and 3D UEGs, which indicates promise for non-trivial results from EDFT on periodic systems.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

DOE OSTI2023

Machine-learning Kohn–Sham potential from dynamics in time-dependent Kohn–Sham systems

Abstract The construction of a better exchange-correlation potential in time-dependent density functional theory (TDDFT) can improve the accuracy of TDDFT calculations and provide more accurate predictions of the properties of many-electron systems. Here, we propose a machine learning method to develop the energy functional and the Kohn–Sham potential of a time-dependent Kohn–Sham (TDKS) system is proposed. The method is based on the dynamics of the Kohn–Sham system and does not require any data on the exact Kohn–Sham potential for training the model. We demonstrate the results of our method with a 1D harmonic oscillator example and a 1D two-electron example. We show that the machine-learned Kohn–Sham potential matches the exact Kohn–Sham potential in the absence of memory effect. Our method can still capture the dynamics of the Kohn–Sham system in the presence of memory effects. The machine learning method developed in this article provides insight into making better approximations of the energy functional and the Kohn–Sham potential in the TDKS system.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2025

Deep Learning without Global Optimization by Random Fourier Neural Networks

Here we introduce a new training algorithm for deep neural networks that utilize random complex exponential activation functions. Our approach employs a Markov chain Monte Carlo sampling procedure to iteratively train network layers, avoiding global and gradient-based optimization while maintaining error control. It consistently attains the theoretical approximation rate for residual networks with complex exponential activation functions, determined by network complexity. Additionally, it enables efficient learning of multiscale and high-frequency features, producing interpretable parameter distributions. Despite using sinusoidal basis functions, we do not observe Gibbs phenomena in approximating discontinuous target functions.

97 MATHEMATICS AND COMPUTING↗

NASA NTRS2011

The Evolution of Ly-alpha Emitting Galaxies Between z = 2.1 and z = 3.l

We describe the results of a new, wide-field survey for z= 3.1 Ly-alpha emission-line galaxies (LAEs) in the Extended Chandra Deep Field South (ECDF-S). By using a nearly top-hat 5010 Angstrom filter and complementary broadband photometry from the MUSYC survey, we identify a complete sample of 141 objects with monochromatic fluxes brighter than 2.4E-17 ergs/cm^2/s and observers-frame equivalent widths greater than ~ 80 Angstroms (i.e., 20 Angstroms in the rest-frame of Ly-alpha). The bright-end of this dataset is dominated by x-ray sources and foreground objects with GALEX detections, but when these interlopers are removed, we are still left with a sample of 130 LAE candidates, 39 of which have spectroscopic confirmations. This sample overlaps the set of objects found in an earlier ECDF-S survey, but due to our filter's redder bandpass, it also includes 68 previously uncataloged sources. We confirm earlier measurements of the z=3.1 LAE emission-line luminosity function, and show that an apparent anti-correlation between equivalent width and continuum brightness is likely due to the effect of correlated errors in our heteroskedastic dataset. Finally, we compare the properties of z=3.1 LAEs to LAEs found at z=2.1. We show that in the approximately 1 Gyr after z approximately 3, the LAE luminosity function evolved significantly, with L * fading by approximately 0.4 mag, the number density of sources with L greater than 1.5E42 ergs/s declining by approximately 50%, and the equivalent width scalelength contracting from 70^{+7}_{-5} Angstroms to 50^{+9}_{-6} Angstroms. When combined with literature results, our observations demonstrate that over the redshift range z approximately 0 to z approximately 4, LAEs contain less than approximately 10% of the star-formation rate density of the universe.

Ciardullo, Robin↗