Search NASA⌕ Search

SEARCH · Search NASA

Results for “Differentiability”

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 901 records · Page 50

Symmetries of stochastic differential equations: A geometric approach

A new notion of stochastic transformation is proposed and applied to the study of both weak and strong symmetries of stochastic differential equations (SDEs). The correspondence between an algebra of weak symmetries for a given SDE and an algebra of strong symmetries for a modified SDE is proved under suitable regularity assumptions. This general approach is applied to a stochastic version of a two dimensional symmetric ordinary differential equation and to the case of two dimensional Brownian motion.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Symmetries of the Gas Dynamics Equations using the Differential Form Method

Here, a brief review of the theory of exterior differential systems and isovector symmetry analysis methods is presented in the context of the one-dimensional inviscid compressible flow equations. These equations are formulated as an exterior differential system with equation of state (EOS) closure provided in terms of an adiabatic bulk modulus. The scaling symmetry generators—and corresponding EOS constraints—otherwise appearing in the existing literature are recovered through the application and invariance under Lie derivative dragging operations.

97 MATHEMATICS AND COMPUTING↗

Learning effective stochastic differential equations from microscopic simulations: Linking stochastic numerics to deep learning

We identify effective stochastic differential equations (SDEs) for coarse observables of fine-grained particle- or agent-based simulations; these SDEs then provide useful coarse surrogate models of the fine scale dynamics. We approximate the drift and diffusivity functions in these effective SDEs through neural networks, which can be thought of as effective stochastic ResNets. The loss function is inspired by, and embodies, the structure of established stochastic numerical integrators (here, Euler–Maruyama and Milstein); our approximations can thus benefit from backward error analysis of these underlying numerical schemes. They also lend themselves naturally to “physics-informed” gray-box identification when approximate coarse models, such as mean field equations, are available. Existing numerical integration schemes for Langevin-type equations and for stochastic partial differential equations can also be used for training; we demonstrate this on a stochastically forced oscillator and the stochastic wave equation. Our approach does not require long trajectories, works on scattered snapshot data, and is designed to naturally handle different time steps per snapshot. We consider both the case where the coarse collective observables are known in advance, as well as the case where they must be found in a data-driven manner.

97 MATHEMATICS AND COMPUTING↗

Parallelizable adjoint stencil computations using transposed forward-mode algorithmic differentiation

Algorithmic differentiation (AD) is a tool for generating discrete adjoint solvers, which efficiently compute gradients of functions with many inputs, for example for use in gradient-based optimization. AD is often applied to large computations such as stencil operators, which are an important part of most structured-mesh PDE solvers. Stencil computations are often parallelized, for example by using OpenMP, and optimized by using techniques such as cache-blocking and tiling to fully utilize multicore CPUs and many-core accelerators and GPUs. Differentiating these codes with conventional reverse-mode AD results in adjoint codes that cannot be expressed as stencil operations and may not be easily parallelizable. They thus leave most of the compute power of modern architectures unused. We present a method that combines forward-mode AD and loop transformation to generate adjoint solvers that use the same memory access pattern as the original computation that they are derived from and can benefit from the same optimization techniques. The effectiveness of this method is demonstrated by generating a scalable adjoint CFD solver for multicore CPUs and Xeon Phi accelerators.

algorithmic differentiation↗

Differential cross sections for excitation of H 2 by low-energy electron impact

Experimental and theoretical differential cross sections (DCS) for the electron-impact excitation of molecular hydrogen to the $B{}^{1}{{\rm{\Sigma }}}_{u}^{+}$, $c{}^{3}{{\rm{\Pi }}}_{u}$, $a{}^{3}{{\rm{\Sigma }}}_{g}^{+}$, $C{}^{1}{{\rm{\Pi }}}_{u}$, and the $E(F){}^{1}{{\rm{\Sigma }}}_{g}^{+}$ states are presented at incident energies near to threshold. The experimental DCSs were taken at incident energies of 14, 15, 16 and 17.5 eV and for scattering angles from 10° to 130°. In conclusion, the theoretical DCSs are from the convergent close-coupling method which has recently successfully modeled differential electron scattering from H 2 when compared with available experiment at energies of 17.5 eV and above.

74 ATOMIC AND MOLECULAR PHYSICS↗

Differentiable physics-enabled closure modeling for Burgers’ turbulence

Abstract Data-driven turbulence modeling is experiencing a surge in interest following algorithmic and hardware developments in the data sciences. We discuss an approach using the differentiable physics paradigm that combines known physics with machine learning to develop closure models for Burgers’ turbulence. We consider the one-dimensional Burgers system as a prototypical test problem for modeling the unresolved terms in advection-dominated turbulence problems. We train a series of models that incorporate varying degrees of physical assumptions on an a posteriori loss function to test the efficacy of models across a range of system parameters, including viscosity, time, and grid resolution. We find that constraining models with inductive biases in the form of partial differential equations that contain known physics or existing closure approaches produces highly data-efficient, accurate, and generalizable models, outperforming state-of-the-art baselines. Addition of structure in the form of physics information also brings a level of interpretability to the models, potentially offering a stepping stone to the future of closure modeling.

97 MATHEMATICS AND COMPUTING↗

Pseudostate methods and differential cross sections for antiproton ionization of atomic hydrogen and helium

A relaxed form of a recent impact parameter coupled pseudostate approximation of McGovern et al. [Phys. Rev. A 79, 042707 (2009)] for calculating differential ionization cross sections is proposed. This greatly eases the computational burden in cases where a range of ejected electron energies has to be considered. The relaxed approximation is tested against exact first Born calculations for antiproton impact on H and nonperturbatively for the highly nonperturbative system of Au{sup 53+} incident upon He. The approximation performs well in these tests. It is shown how, with a little further approximation, the relaxed theory leads to a widely used prescription for the total ionization cross section. Results for differential ionization of H and He by antiprotons are presented. These reveal the growing dominance of the interaction between the antiproton and the target nucleus at low impact energies and show the changing importance of the role of the postcollisional interaction between the antiproton and the ejected electron.

74 ATOMIC AND MOLECULAR PHYSICS↗

Low-momentum-transfer nonrelativistic limit of the relativistic impulse approximation expression for Compton-scattering doubly differential cross sections and characterization of their relativistic contributions

The nonrelativistic (nr) impulse approximation (NRIA) expression for Compton-scattering doubly differential cross sections (DDCS) for inelastic photon scattering is recovered from the corresponding relativistic expression (RIA) of Ribberfors [Phys. Rev. B 12, 2067 (1975)] in the limit of low momentum transfer (q{yields}0), valid even at relativistic incident photon energies {omega}{sub 1}>m provided that the average initial momentum of the ejected electron <p{sub i}> is not too high, that is, <p{sub i}> <m. This corresponds to a binding energy E{sub b}<10 keV. This q{yields}0 nr limit is simultaneous with the approach of the scattering angle {theta} to 0 deg. ({theta}{yields}0 deg.) around the Compton peak maximum. This explains the observation that it is possible to obtain an accurate Compton peak (CP) even when {omega}{sub 1}>m using nr expressions when {theta} is small. For example, a 1% accuracy can be obtained when {omega}{sub 1}=1 MeV if {theta}<20 deg. However as {omega}{sub 1} increases into the MeV range, the maximum {theta} at which an accurate Compton peak can be obtained from nr expressions approaches closer to zero, because the {theta} at which the relativistic shift of CP to higher energy is greatest, which starts at 180 deg. when {omega}{sub 1}<300 keV, begins to decrease, approaching zero even though the {theta} at which the relativistic increase in the CP magnitude remains greatest around {theta}=180 deg. The relativistic contribution to the prediction of Compton doubly differential cross sections (DDCS) is characterized in simple terms using Ribberfors further approximation to his full RIA expression. This factorable form is given by DDCS=KJ, where K is the kinematic factor and J the Compton profile. This form makes it possible to account for the relativistic shift of CP to higher energy and the increase in the CP magnitude as being due to the dependence of J(p{sub min},{rho}{sub rel}) (where p{sub min} is the relativistic version of the z component of the momentum of the initial electron and {rho}{sub rel} is the relativistic charge density) and K(p{sub min}) on p{sub min}. This characterization approach was used as a guide for making the nr QED S-matrix expression for the Compton peak kinematically relativistic. Such modified nr expressions can be more readily applied to large systems than the fully relativistic version.

74 ATOMIC AND MOLECULAR PHYSICS↗

Resonant vibrational excitation of H{sub 2} by electron impact: Full-range differential cross sections

Electron-impact vibrational excitation of the hydrogen molecule has been revisited in the energy region from 1 to 5 eV. A crossed-beam double trochoidal electron spectrometer is used. Forward and backward scattered electrons from the v=0{yields}1 excitation channel are separated by electron-beam modulation and time-of-flight detection technique. Present results are normalized and absolute values of differential cross sections at critical border angles of 0 deg. and 180 deg. are determined. In this way the differential cross-section measurements are completed in the full angular range from 0 deg. to 180 deg.

74 ATOMIC AND MOLECULAR PHYSICS↗

Bethe binary-encounter peaks in the double-differential cross sections for high-energy electron-impact ionization of H{sub 2} and He

We study the Bethe binary-encounter (BE) region in the ejected-electron double-differential emission spectrum of H{sub 2} and He targets in collisions with 8-keV electrons. We compare the absolute cross sections for these isoelectronic systems at high emission energies. The experimental data are analyzed in terms of a state-of-the-art theoretical model based on a two-effective-center approximation. In the case of the H{sub 2} molecule the binary peak in the double-differential cross sections (DDCS) is enhanced due to the two-center Young-type interference. The observed undulation in the DDCS ratio is explained in terms of the combined contributions of the Compton profile mismatch and the interference effect. The influence of the interference effect is thus observed for higher-energy electrons compared to most of the earlier studies which focused on low-energy electrons produced in soft collisions.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Near-threshold electron-impact doubly differential cross sections for the ionization of argon and krypton

We present normalized doubly differential cross sections (DDCS's) for the near-threshold, electron-impact single ionization of argon and krypton, similar to those taken earlier for Ne and Xe [Yates et al., J. Phys. B 42, 095206 (2009)]. The Ar measurements were taken at incident energies of 17, 18, 20, and 30 eV while the Kr measurements were taken at 15, 16, 17.5, and 20 eV. The DDCS scattering angles range from 15 deg. to 120 deg. The differential data are initially normalized to available experimental cross sections for excitation of the ground np{sup 6} to the np{sup 5}(n+1)s excited states of the noble gas and, after integration, to well-established experimental total ionization cross sections of Rapp and Englander-Golden [J. Chem. Phys. 43, 1464 (1965)].

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Electron-quasihole duality and second-order differential equation for Read-Rezayi and Jack wave functions

We consider the quasihole wave functions of the non-Abelian Read-Rezayi quantum-Hall states which are given by the conformal blocks of the minimal model WA{sub k-1}(k+1,k+2) of the WA{sub k-1} algebra. By studying the degenerate representations of this conformal field theories, we derive a second-order differential equation satisfied by a general many-quasihole wave function. We find a duality between the differential equations fixing the electron and quasihole wave functions. They both satisfy the Laplace-Beltrami equation. We use this equation to obtain an analytic expression for the generic wave function with one excess flux. These results also apply to the more general models WA{sub k-1}(k+1,k+r) corresponding to the recently introduced Jack states.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Differential cross sections of the charge-exchange reaction {pi}{sup -}p{yields}{pi}{sup 0}n in the momentum range from 103 to 178 MeV/c

Measured values of the differential cross sections for pion-nucleon charge exchange, {pi}{sup -}p{yields}{pi}{sup 0}n, are presented for {pi}{sup -} momenta of 103,112,120,130,139,152, and 178 MeV/c. Complete angular distributions were obtained by using the Crystal Ball detector at the Alternating Gradient Synchrotron at Brookhaven National Laboratory. Statistical uncertainties of the differential cross sections vary from 3 to 6% in the backward angle region and from 6 to about 20% in the forward region with the exception of the two most forward angles. The systematic uncertainties are estimated to be about 3% for all momenta.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Predicted levels of {sup 9}Be based on a theoretical analysis of neutron double-differential cross sections at E{sub n}=14.1 and 18 MeV

By using the statistical theory for neutron-induced light nucleus reaction, the calculation of the neutron double-differential cross sections for n+{sup 9}Be reactions is performed. The secondary outgoing neutrons only coming from the (n,2n)2{alpha} reaction channel through six different emission processes are illustrated in detail in this article. Based on the theoretical analysis of neutron double-differential cross sections at E{sub n}=14.1 and 18 MeV, two predicted levels of {sup 9}Be, i.e., E(J{sup {pi}}){gamma}=9((5/2){sup +})1000 and 10((5/2){sup +})1000, have been recommended. The calculated results indicate that the fittings would be improved obviously while the predicted levels have been employed.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

First measurement of the muon neutrino charged current quasielastic double differential cross section

A high-statistics sample of charged-current muon neutrino scattering events collected with the MiniBooNE experiment is analyzed to extract the first measurement of the double differential cross section ((d{sup 2{sigma}}/dT{sub {mu}d}cos{theta}{sub {mu}})) for charged-current quasielastic (CCQE) scattering on carbon. This result features minimal model dependence and provides the most complete information on this process to date. With the assumption of CCQE scattering, the absolute cross section as a function of neutrino energy ({sigma}[E{sub {nu}}]) and the single differential cross section ((d{sigma}/dQ{sup 2})) are extracted to facilitate comparison with previous measurements. These quantities may be used to characterize an effective axial-vector form factor of the nucleon and to improve the modeling of low-energy neutrino interactions on nuclear targets. The results are relevant for experiments searching for neutrino oscillations.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Forward-backward asymmetry and the differential cross section of the top quark in the flavor violating Z{sup '} model at O({alpha}{sub s}{sup 2{alpha}}{sub X})

In this paper, forward-backward asymmetry and differential cross section of top quark in flavor violating Z{sup '} model up to O({alpha}{sub s}{sup 2{alpha}}{sub X}) at Tevatron are calculated. In order to account for the top observed large forward-backward asymmetry, the new coupling g{sub X} among Z{sup '} and quarks will be not much less than strong coupling constant g{sub s}. After including the new higher order correction, the differential cross section can fit the data better than those only including the leading contributions from Z{sup '}, while the forward-backward asymmetry is still in agreement with the measurement.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Alignment of low-dose X-ray fluorescence tomography images using differential phase contrast

X-ray fluorescence nanotomography provides unprecedented sensitivity for studies of trace metal distributions in whole biological cells. Dose fractionation, in which one acquires very low dose individual projections and then obtains high statistics reconstructions as signal from a voxel is brought together (Hegerl & Hoppe, 1976), requires accurate alignment of these individual projections so as to correct for rotation stage runout. It is shown here that differential phase contrast at 10.2 keV beam energy offers the potential for accurate cross-correlation alignment of successive projections, by demonstrating that successive low dose, 3 ms per pixel, images acquired at the same specimen position and rotation angle have a narrower and smoother cross-correlation function (1.5 pixels FWHM at 300 nm pixel size) than that obtained from zinc fluorescence images (25 pixels FWHM). The differential phase contrast alignment resolution is thus well below the 700 nm × 500 nm beam spot size used in this demonstration, so that dose fractionation should be possible for reduced-dose, more rapidly acquired, fluorescence nanotomography experiments.

differential phase contrast↗

MapsTorch : automatic differentiation for X-ray fluorescence data analysis

X-ray fluorescence (XRF) is a popular spectroscopy technique for elemental analysis. Spectrum fitting and parameter tuning are at the core of XRF analysis and are conventionally manually intensive, especially for synchrotron experiments involving large amounts of diverse samples. This work introduces the automatic differentiation (AD) technique to XRF and an open-source package called MapsTorch. By transforming an analytical model of the XRF spectrum into a differentiable computation graph with AD, MapsTorch enables robust optimization of parameters and elemental intensities. We evaluate MapsTorch by conducting computational experiments on a large number of historical synchrotron XRF datasets and compare its performance with the currently practiced fitting tool NLopt. The results show that MapsTorch consistently achieves high-quality fits and often leads to better fitting quality than NLopt, particularly in tasks such as initial spectrum fitting and elemental intensity refinement. The robust performance of MapsTorch paves the way for developing automated and high-throughput XRF data analysis workflows to handle the increasing data volumes expected from next-generation synchrotron facilities.

X-ray fluorescence↗