Search NASA⌕ Search

SEARCH · Search NASA

Results for “tensor methods”

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 991 records · Page 55

Upscaling permeability for three-dimensional fractured porous rocks with the multiple boundary method

Upscaling permeability of grid blocks is crucial for groundwater models. A novel upscaling method for three-dimensional fractured porous rocks is presented. The objective of the study was to compare this method with the commonly used Oda upscaling method and the volume averaging method. First, the multiple boundary method and its computational framework were defined for three-dimensional stochastic fracture networks. Then, the different upscaling methods were compared for a set of rotated fractures, for tortuous fractures, and for two discrete fracture networks. The results computed by the multiple boundary method are comparable with those of the other two methods and fit best the analytical solution for a set of rotated fractures. The errors in flow rate of the equivalent fracture model decrease when using the multiple boundary method. Furthermore, the errors of the equivalent fracture models increase from well-connected fracture networks to poorly connected ones. Finally, the diagonal components of the equivalent permeability tensors tend to follow a normal or log-normal distribution for the well-connected fracture network model with infinite fracture size. By contrast, they exhibit a power-law distribution for the poorly connected fracture network with multiple scale fractures. The study demonstrates the accuracy and the flexibility of the multiple boundary upscaling concept. This makes it attractive for being incorporated into any existing flow-based upscaling procedures, which helps in reducing the uncertainty of groundwater models.

58 GEOSCIENCES↗

DeepCMB: Lensing reconstruction of the cosmic microwave background with deep neural networks

Next-generation cosmic microwave background (CMB) experiments will have lower noise and therefore increased sensitivity, enabling improved constraints on fundamental physics parameters such as the sum of neutrino masses and the tensor-to-scalar ratio r. Achieving competitive constraints on these parameters requires high signal-to-noise extraction of the projected gravitational potential from the CMB maps. Standard methods for reconstructing the lensing potential employ the quadratic estimator (QE). However, the QE performs suboptimally at the low noise levels expected in upcoming experiments. Other methods, like maximum likelihood estimators (MLE), are under active development. In this work, we demonstrate reconstruction of the CMB lensing potential with deep convolutional neural networks (CNN) - ie, a ResUNet. The network is trained and tested on simulated data, and otherwise has no physical parametrization related to the physical processes of the CMB and gravitational lensing. We show that, over a wide range of angular scales, ResUNets recover the input gravitational potential with a higher signal-to-noise ratio than the QE method, reaching levels comparable to analytic approximations of MLE methods. We demonstrate that the network outputs quantifiably different lensing maps when given input CMB maps generated with different cosmologies. We also show we can use the reconstructed lensing map for cosmological parameter estimation. This application of CNN provides a few innovations at the intersection of cosmology and machine learning. First, while training and regressing on images, we predict a continuous-variable field rather than discrete classes. Second, we are able to establish uncertainty measures for the network output that are analogous to standard methods. We expect this approach to excel in capturing hard-to-model non-Gaussian astrophysical foreground and noise contributions.

79 ASTRONOMY AND ASTROPHYSICS↗

Fluctuations and phase transitions of uniaxial and biaxial liquid crystals using a theoretically informed Monte Carlo and a Landau free energy density

In this work, we explore fluctuations during phase transitions of uniaxial and biaxial liquid crystals using a phenomenological free energy functional. We rely on a continuum-level description of the liquid crystal ordering with a tensorial parameter and a temperature dependent Landau polynomial expansion of the tensor's invariants. The free energy functional, over a three-dimensional periodic domain, is integrated with a Gaussian quadrature and minimized with a theoretically informed Monte Carlo method. We reconstruct analytical phase diagrams, following Landau and Doi 's notations, to verify that the free energy relaxation reaches the global minimum. Importantly, our relaxation method is able to follow the thermodynamic behavior provided by other non-phenomenological approaches; we predict the first order character of the isotropic-nematic transition, and we identify the uniaxial-biaxial transition as second order. Finally, we use a finite-size scaling method, using the nematic susceptibility, to calculate the transition temperatures for 4-Cyano-4'-pentylbiphenyl (5CB) and N-(4-methoxybenzylidene)-4-butylaniline (MBBA). Our results show good agreement with experimental values, thereby validating our minimization method. Our approach is an alternative towards the relaxation of temperature dependent continuum-level free energy functionals, in any geometry, and can incorporate complicated elastic and surface energy densities.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Mixeds-sourcery: Building many-body states using bubbles of nothing

We recently introduced the idea of s-sourcery [B. Swingle and J. McGreevy, Phys. Rev. B 93, 045127 (2016)], a general formalism for building many-body quantum ground states using renormalization-group-inspired quantum circuits. In this work, we define a generalized notion of s-sourcery that applies to mixed states and study its properties and applicability. For our examples we focus on thermal states of local Hamiltonians. We prove a number of theorems establishing the prevalence of mixed s-source fixed points, giving results for free fermion models, conformal field theories, holographic models, and topological phases. Thermal double states (also called thermofield double states) and the machinery of approximate conditional independence are used heavily in the constructions. For a large class of models we provide an information theoretic argument for the existence of a local Hamiltonian whose ground state is the thermal double state, and in some cases we construct such a Hamiltonian.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Scalar-tensor cosmologies: Fixed points of the Jordan frame scalar field

We study the evolution of homogeneous and isotropic, flat cosmological models within the general scalar-tensor theory of gravity with arbitrary coupling function and potential. After introducing the limit of general relativity we describe the details of the phase space geometry. Using the methods of dynamical systems for the decoupled equation of the Jordan frame scalar field we find the fixed points of flows in two cases: potential domination and matter domination. We present the conditions on the mathematical form of the coupling function and potential which determine the nature of the fixed points (attractor or other). There are two types of fixed points, both are characterized by cosmological evolution mimicking general relativity, but only one of the types is compatible with the Solar System parametrized post-Newtonian (PPN) constraints. The phase space structure should also carry over to the Einstein frame as long as the transformation between the frames is regular which however is not the case for the latter (PPN compatible) fixed point.

79 ASTRONOMY AND ASTROPHYSICS↗

Weak chaos and the 'melting transition' in a confined microplasma system

We present results demonstrating the occurrence of changes in the collective dynamics of a Hamiltonian system which describes a confined microplasma characterized by long-range Coulomb interactions. In its lower energy regime, we first detect macroscopically the transition from a 'crystallinelike' to a 'liquidlike' behavior, which we call the 'melting transition'. We then proceed to study this transition using a microscopic chaos indicator called the smaller alignment index (SALI), which utilizes two deviation vectors in the tangent dynamics of the flow and is nearly constant for ordered (quasiperiodic) orbits, while it decays exponentially to zero for chaotic orbits as exp[-(lambda{sub 1}-lambda{sub 2})t], where lambda{sub 1}>lambda{sub 2}>0 are the two largest Lyapunov exponents. During the melting phase, SALI exhibits a peculiar stairlike decay to zero, reminiscent of 'sticky' orbits of Hamiltonian systems near the boundaries of resonance islands. This alerts us to the importance of the DELTAlambda=lambda{sub 1}-lambda{sub 2} variations in that regime and helps us identify the energy range over which 'melting' occurs as a multistage diffusion process through weakly chaotic layers in the phase space of the microplasma. Additional evidence supporting further the above findings is given by examining the GALI{sub k} indices, which generalize SALI (=GALI{sub 2}) to the case of k>2 deviation vectors and depend on the complete spectrum of Lyapunov exponents of the tangent flow about the reference orbit.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Ex-Situ Stress Measurements in Polycrystalline Ceramics Using Photo-Stimulated Luminescence Spectroscopy and High-Energy X-Rays

The photo-stimulated luminescence spectroscopy (PSLS) technique provides the means to establish stress dependencies of the well-known R-line peak positions in polycrystalline alumina. The uniaxial compression tests presented in this paper, which determine the coefficients describing this piezospectroscopic (PS) behavior, tackle two previously unexplored areas. Firstly, the vibronic band peaks in the emission spectrum are introduced here, and the PS coefficients of several peaks within these bands are established. These results set the foundation for the exploitation of vibronic band peakshifts, along with the R-lines, in order to provide the non-symmetric components of the stress tensor and therefore the measurement of the complete stress state in polycrystalline alumina. Secondly, high-energy X-rays serve as an ex-situ stress measurement method upon which the optical fluorescence shifts are based, thereby advancing the accuracy of PS coefficient determination in a unique approach. The results of PSLS and synchrotron X-ray experiments are presented here and used in conjunction to reveal new information on the PS behavior of polycrystalline alumina.

43 PARTICLE ACCELERATORS↗

Cerebral perfusion computed tomography deconvolution via structure tensor total variation regularization

Purpose: Cerebral perfusion computed tomography (PCT) imaging as an accurate and fast acute ischemic stroke examination has been widely used in clinic. Meanwhile, a major drawback of PCT imaging is the high radiation dose due to its dynamic scan protocol. The purpose of this work is to develop a robust perfusion deconvolution approach via structure tensor total variation (STV) regularization (PD-STV) for estimating an accurate residue function in PCT imaging with the low-milliampere-seconds (low-mAs) data acquisition. Methods: Besides modeling the spatio-temporal structure information of PCT data, the STV regularization of the present PD-STV approach can utilize the higher order derivatives of the residue function to enhance denoising performance. To minimize the objective function, the authors propose an effective iterative algorithm with a shrinkage/thresholding scheme. A simulation study on a digital brain perfusion phantom and a clinical study on an old infarction patient were conducted to validate and evaluate the performance of the present PD-STV approach. Results: In the digital phantom study, visual inspection and quantitative metrics (i.e., the normalized mean square error, the peak signal-to-noise ratio, and the universal quality index) assessments demonstrated that the PD-STV approach outperformed other existing approaches in terms of the performance of noise-induced artifacts reduction and accurate perfusion hemodynamic maps (PHM) estimation. In the patient data study, the present PD-STV approach could yield accurate PHM estimation with several noticeable gains over other existing approaches in terms of visual inspection and correlation analysis. Conclusions: This study demonstrated the feasibility and efficacy of the present PD-STV approach in utilizing STV regularization to improve the accuracy of residue function estimation of cerebral PCT imaging in the case of low-mAs.

60 APPLIED LIFE SCIENCES↗

Isotropic and anisotropic dark energy models

In this review we discuss the evolution of the universe filled with dark energy with or without perfect fluid. In doing so we consider a number of cosmological models, namely Bianchi type I, III, V, VI{sub 0}, VI and FRW ones. For the anisotropic cosmological models we have used proportionality condition as an additional constrain. The exact solutions to the field equations in quadrature are found in case of a BVI model. It was found that the proportionality condition used here imposed severe restriction on the energy-momentum tensor, namely it leads to isotropic distribution of matter. Anisotropic BVI{sub 0}, BV, BIII and BIDE models with variable EoS parameter ω have been investigated by using a law of variation for the Hubble parameter. In this case the matter distribution remains anisotropic, though depending on the concrete model there appear different restrictions on the components of energy-momentum tensor. That is why we need an extra assumption such as variational a law for the Hubble parameter. It is observed that, at the early stage, the EoS parameter v is positive i.e. the universe was matter dominated at the early stage but at later time, the universe is evolving with negative values, i.e., the present epoch. DE model presents the dynamics of EoS parameter ω whose range is in good agreement with the acceptable range by the recent observations. A spatially homogeneous and anisotropic locally rotationally symmetric Bianchi-I space time filled with perfect fluid and anisotropic DE possessing dynamical energy density is studied. In the derived model, the EoS parameter of DE (ω{sup (de)}) is obtained as time varying and it is evolving with negative sign which may be attributed to the current accelerated expansion of Universe. The distance modulus curve of derived model is in good agreement with SNLS type Ia supernovae for high redshift value which in turn implies that the derived model is physically realistic. A system of two fluids within the scope of a spatially flat and isotropic FRW model is studied. The role of the two fluids, either minimally or directly coupled in the evolution of the dark energy parameter, has been investigated. In doing so we have used three different ansatzs regarding the scale factor that gives rise to a variable decelerating parameter. It is observed that, in the non-interacting case, both the open and flat universes can cross the phantom region whereas in the interacting case only the open universe can cross the phantom region. The stability and acceptability of the obtained solution are also investigated.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Apparent Explosion Moments from Rg Waves Recorded on SPE

Seismic moments for the first four chemical tests making up phase I of the Source Physics Experiments (SPE) are estimated from 6-Hz Rg waves recorded along a single radial line of geophones under the assumption that the tests are pure explosions. These apparent explosion moments are compared with moments determined from the reduced displacement potential method applied to free-field data. Light detection and ranging (lidar) observations, strong ground motions on the free surface in the vicinity of ground zero, and moment tensor inversion results are evidence that the fourth test SPE-4P is a pure explosion, and the moments show good agreement, 8×10 10 N·m for free-field data versus 9×10 10 N·m for Rg waves. In stark contrast, apparent moments for the first three tests are smaller than near-field moments by factors of 3–4. Relative amplitudes for the three tests determined from Rg interferometry using SPE-4P as an empirical Green’s function indicate that radiation patterns are cylindrically symmetric within a factor of 1.25 (25%). This fact assures that the apparent moments are reliable even though they were measured on just one azimuth. Spallation occurred on the first three tests, and ground-based lidar detected permanent deformations. As such, the source medium suffered late-time damage. In conclusion, destructive interference between Rg waves radiated by explosion and damage sources will reduce amplitudes and explain why apparent moments are smaller than near-field moments based on compressional energy emitted directly from the source.

47 OTHER INSTRUMENTATION↗

DeepCMB: Lensing Reconstruction of the CMB with Deep Neural Networks

Next-generation cosmic microwave background (CMB) experiments will have lower noise and therefore increased sensitivity, enabling improved constraints on fundamental physics parameters such as the sum of neutrino masses and the tensor-to-scalar ratio r. Achieving competitive constraints on these parameters requires high signal-to-noise extraction of the projected gravitational potential from the CMB maps. Standard methods for reconstructing the lensing potential employ the quadratic estimator (QE). However, the QE performs suboptimally at the low noise levels expected in upcoming experiments. Other methods, like maximum likelihood estimators (MLE), are under active development. In this work, we demonstrate reconstruction of the CMB lensing potential with deep convolutional neural networks (CNN) - ie, a ResUNet. The network is trained and tested on simulated data, and otherwise has no physical parametrization related to the physical processes of the CMB and gravitati onal lensing. We show that, over a wide range of angular scales, ResUNets recover the input gravitational potential with a higher signal-to-noise ratio than the QE method, reaching levels comparable to analytic approximations of MLE methods. We demonstrate that the network outputs quantifiably different lensing maps when given input CMB maps generated with different cosmologies. We also show we can use the reconstructed lensing map for cosmological parameter estimation. This application of CNN provides a few innovations at the intersection of cosmology and machine learning. First, while training and regressing on images, we predict a continuous-variable field rather than discrete classes. Second, we are able to establish uncertainty measures for the network output that are analogous to standard methods. We expect this approach to excel in capturing hard-to-model non-Gaussian astrophysical foreground and noise contributions.

79 ASTRONOMY AND ASTROPHYSICS↗

Static versioning in the polyhedral model

An approach is presented to enhancing the optimization process in a polyhedral compiler by introducing compile-time versioning, i.e., the production of several versions of optimized code under varying assumptions on its run-time parameters. We illustrate this process by enabling versioning in the polyhedral processor placement pass. We propose an efficient code generation method and validate that versioning can be useful in a polyhedral compiler by performing benchmarking on a small set of deep learning layers defined for dynamically-sized tensors.

Meister, Benoit J.↗

The Polarization Tensor of the Massless Mode in Yang-Mills Thermodynamics

We discuss the constraints on energy-momentum transfers in local vertices in the effective theory for the deconfining phase of SU(2) Yang-Mills thermodynamics. Subsequently, we apply these constraints to the computation of the polarization tensor of the massless mode on the one-loop level. The resulting gap equation for the sought-after screening function is solved numerically. We discuss and interpret our results.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Isovector and isoscalar tensor charges of the nucleon from lattice QCD

We present results for the iso-vector and flavor diagonal tensor charges $g$$^{u−d}_{T}$, $g$$^{u}_{T}$, $g$$^{d}_{T}$, and $g$$^{s}_{T}$ needed to probe novel tensor interactions at the TeV scale in neutron and nuclear β-decays and the contribution of the quark electric dipole moment (EDM) to the neutron EDM. The lattice QCD calculations were done using nine ensembles of gauge configurations generated by the MILC collaboration using the HISQ action with 2+1+1 dynamical flavors. These ensembles span three lattice spacings a ≈ 0.06, 0.09 and 0.12 fm and three quark masses corresponding to the pion masses M π ≈ 130, 220 and 310 MeV. Using estimates from these ensembles, we quantify all systematic uncertainties and perform a simultaneous extrapolation in the lattice spacing, volume and light quark masses for the connected contributions. The final estimates of the connected nucleon (proton) tensor charge for the iso-vector combination is $g$$^{u−d}_{T}$ = 1.020(76) in the MS scheme at 2 GeV. The additional disconnected quark loop contributions needed for the flavor-diagonal matrix elements are calculated using a stochastic estimator employing the truncated solver method with the all-mode-averaging technique. We find that the size of the disconnected contribution is smaller than the statistical error in the connected contribution. This allows us to bound the disconnected contribution and include it as an additional uncertainty in the flavor-diagonal charges. After a continuum extrapolation, we find $g$$^{u}_{T}$ = 0.774(66), $g$$^{d}_{T}$ = −0.233(28) and $g$$^{u+d}_{T}$ = 0.541(67). The strangeness tensor charge, that can make a significant contribution to the neutron EDM due to the large ratio ms/mu,d, is $g$$^{s}_{T}$ = 0.008(9) in the continuum limit.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Spectrally Stabilized Interface Capturing Formulation and Implementation in Nek5000/NekRS

This report documents the formulation of a novel level-set method for incompressible two-phase flows in the continuous Galerkin (CG) high order spectral element framework. The overall method hinges on a novel implementation of the spectral vanishing viscosity (SVV) operator for the stabilization of linear/non-linear hyperbolic problems. The multidimensional SVV convolution kernels, which in essence, have a similar effect as a high pass filter applied to the derivatives, are formulated by exploiting the tensor product form, analogous to the construction of the usual stiffness matrix system. The resulting kernels are directionally decoupled and ensure a linear, symmetric positive definite, elliptic matrix operator. The SVV formulation is demonstrated to provide a robust stabilizing mechanism through challenging linear and non-linear hyperbolic problems, including problems pertinent to the level-set formulation. The two-phase framework conceptualized herein is based on the conservative level-set (CLS) method which represents the interface between the fluids by the 0.5 iso-contour of the smoothed Heaviside function. The CLS method is augmented with a preconditioning procedure for interface normals using the signed distance function which precludes the manifestation of spurious oscillations in the vicinty of the interface. Further, the existing mixed explicit-implicit approach for the solution of Navier-Stokes equations in Nek5000, as described in Tomboulides et al, is augmented with a pressure coefficient splitting approach for the Poisson equation, which greatly accelerated the convergence of pressure solver for two-phase systems with large density ratio. The robustness and accuracy of the overall two-phase method is demonstrated through canonical challenging problems involving high density and viscosity ratios, with and without surface tension. The two-phase formulation is wholly implemented in Nek5000 and the SVV stabilization method is implemented in NekRS, which is the essential precursor to the two-phase framework, undergoing active development.

97 MATHEMATICS AND COMPUTING↗

One-loop corrections to the power spectrum in general single-field inflation

We perfom a thorough computation of the one-loop corrections from both scalar and tensor degrees of freedom to the power spectrum of curvature fluctuations for non-canonical Lagrangians in single-field inflation. We consider models characterized by a small sound speed c{sub s}, which produce large non-Gaussianities. As expected, the corrections turn out to be inversely proportional to powers of c{sub s}; evaluating their amplitudes it is then possible to derive some theoretical bounds on the sound speed by requesting the conditions necessary for perturbation theory to hold.

79 ASTRONOMY AND ASTROPHYSICS↗

AdS{sub 3} backgrounds from 10D effective action of heterotic string theory

We present a method for calculating solutions and corresponding central charges for backgrounds with AdS{sub 3} and S{sup k} factors in {alpha}{sup '}-exact fashion from the full tree-level low-energy effective action of heterotic string theory. Three examples are explicitly presented: AdS{sub 3}xS{sup 3}xT{sup 4}, AdS{sub 3}xS{sup 2}xS{sup 1}xT{sup 4}, and AdS{sub 3}xS{sup 3}xS{sup 3}xS{sup 1}. Crucial property which enabled our analysis is vanishing of the Riemann tensor calculated from connection with ''{sigma}-model torsion.'' We show the following: (i) Chern-Simons terms are the only source of {alpha}{sup '} corrections not only in BPS, but also in non-BPS cases, suggesting a possible extension of general method of Kraus and Larsen, (ii) our results are in agreement with some conjectures on the form of the part of tree-level Lagrangian not connected to a mixed Chern-Simons term by supersymmetry (and present in all supersymmetric string theories), (iii) new {alpha}{sup '}-exact result for central charges in AdS{sub 3}xS{sup 3}xS{sup 3}xS{sup 1} geometry. As a tool we used our generalization of Sen's E-function formalism to AdS{sub p} with p>2, and paid special attention to proper definition of asymptotic charges.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

How to upload a physical quantum state into correlation space

In the framework of the computational tensor network [Phys. Rev. Lett. 98, 220503 (2007)], the quantum computation is performed in a virtual linear space called the correlation space. It was recently shown [Phys. Rev. Lett. 103, 050503 (2009)] that a state in a correlation space can be downloaded to the real physical space. In this paper, conversely, we study how to upload a state from a real physical space to the correlation space. After showing the impossibility of cloning a state between a real physical space and the correlation space, we propose a simple teleportation-like method of uploading. This method also enables the Gottesman-Chuang gate teleportation trick and entanglement swapping in the virtual-real hybrid setting. Furthermore, compared with the inverse of the downloading method by Cai et al. [Phys. Rev. Lett. 103, 050503 (2009)], which also works to upload, the proposed uploading method has several advantages.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗