Search NASA⌕ Search

SEARCH · Search NASA

Results for “randomized matrix 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.

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

At least 127 records · Page 7

Efficient Unitary Designs from Random Sums and Permutations

A unitary k-design is an ensemble of unitaries that matches the first k moments of the Haar measure. In this work, we provide two efficient constructions of k-designs on n-qubits using new random matrix theory techniques. Our first construction is based on exponentiating sums of random i.i.d. Hermitian matrices and uses O(k2n2)-many gates. In the spirit of central limit theorems, we show that this random sum approximates the Gaussian Unitary Ensemble (GUE). We then show that the product of just two exponentiated GUE matrices is already approximately Haar random. Our second construction is based on products of exponentiated sums of random permutations and uses Õ(k poly (n)) many gates. The k dependence is optimal (up to polylogarithmic factors) and is inherited from the efficiency of existing k-wise independent permutations. Furthermore, replacing random permutations with quantum-secure pseudorandom permutations (PRPs), we also obtain a pseudorandom unitary (PRU) ensemble that is secure under nonadaptive queries. A central feature of both proofs is a new connection between the polynomial method in quantum query complexity and the large-dimension (N) expansion in random matrix theory. In particular, the first construction uses the polynomial method to control high moments of certain random matrix ensembles without requiring delicate Weingarten calculations. In doing so, we define and solve a moment problem on the unit circle, asking whether a finite number of equally weighted points can reproduce a given set of moments. In our second construction, the key step is to exhibit an orthonormal basis for irreducible representations of the partition algebra that has a low-degree large-N expansion. This allows us to show that the distinguishing probability is a low-degree rational polynomial of the dimension N.

algebra↗

Randomized algorithms for accelerating linear algebraic computations

The project supported the development of new methodologies for performing matrix computations that form key building blocks in modern scientific computing, such as low rank approximation of matrices, and efficient representations of global operators that arise in simulations of physical phenomena.

97 MATHEMATICS AND COMPUTING↗

Randomized low-rank decompositions of nuclear three-body interactions

First-principles simulations of many-fermion systems are commonly limited by the computational requirements of processing large data objects. As a remedy, we propose the use of low-rank approximations of three-body interactions, which are the dominant such limitation in nuclear physics. We introduce a randomized decomposition technique to handle the excessively large matrix dimensions and study the sensitivity of low-rank properties to interaction details. The developed low-rank three-nucleon interactions are benchmarked in ab initio simulations of few- and many-body systems. Exploiting low-rank properties provides a promising route to extend the microscopic description of atomic nuclei to large systems where storage requirements exceed the computational capacities of the most advanced high-performance computing facilities.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Vector solution for the mean electromagnetic fields in a layer of random particles

The mean electromagnetic fields are found in a layer of randomly oriented particles lying over a half space. A matrix-dyadic formulation of Maxwell's equations is employed in conjunction with the Foldy-Lax approximation to obtain equations for the mean fields. A two variable perturbation procedure, valid in the limit of small fractional volume, is then used to derive uncoupled equations for the slowly varying amplitudes of the mean wave. These equations are solved to obtain explicit expressions for the mean electromagnetic fields in the slab region in the general case of arbitrarily oriented particles and arbitrary polarization of the incident radiation. Numerical examples are given for the application to remote sensing of vegetation.

Lang, R. H.↗

Doorway states in the random-phase approximation

By coupling a doorway state to a sea of random background states, we develop the theory of doorway states in the framework of the random-phase approximation (RPA). Because of the symmetry of the RPA equations, that theory is radically different from the standard description of doorway states in the shell model. We derive the Pastur equation in the limit of large matrix dimension and show that the results agree with those of matrix diagonalization in large spaces. The complexity of the Pastur equation does not allow for an analytical approach that would approximately describe the doorway state. Our numerical results display unexpected features: The coupling of the doorway state with states of opposite energy leads to strong mutual attraction.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

A Galerkin method with two-dimensional Haar basis functions for the computation of the Karhunen–Loève expansion

We study the numerical approximation of a homogeneous Fredholm integral equation of second kind associated with the Karhunen–Loève expansion of Gaussian random fields. We develop, validate, and discuss an algorithm based on the Galerkin method with two-dimensional Haar wavelets as basis functions. The shape functions are constructed from the orthogonal decomposition of tensor-product spaces of one-dimensional Haar functions, and a recursive algorithm is employed to compute the matrix of the discrete eigenvalue system without the explicit calculation of integrals, allowing the implementation of a fast and efficient algorithm that provides considerable reduction in CPU time, when compared with classical Galerkin methods. Numerical experiments confirm the convergence rate of the method and assess the approximation error and the sparsity of the eigenvalue system when the wavelet expansion is truncated. We illustrate the numerical solution of a diffusion problem with random input data with the present method. In this problem, accuracy was retained after dropping the coefficients below a threshold value that was numerically determined. A similar method with scaling functions rather than wavelet functions does not need a discrete wavelet transform and leads to eigenvalue systems with better conditioning but lower sparsity.

97 MATHEMATICS AND COMPUTING↗

Charge-changing particle-hole excitation of {sup 16}N and {sup 16}F nuclei

The nuclear structure of {sup 16}N (closed shell + ν − π) and {sup 16}F (closed shell + π − ν) nuclei is studied using particle-hole proton-neutron Tamm-Dancoff Approximation pn TDA and particle-hole proton-neutron Random Phase Approximation pn RPA. The particle-hole Hamiltonian of PSD model space is to be diagonalized in the presence of the PSDMWKPN interaction: for P-space the Cohen-Kurath interaction, for SD-space the Wildenthal Interaction, for the coupling matrix elements between P- and SD-spaces the Millener-Kurath interaction is used, spurious components were eliminated with CM contribution. The results containing energy level schemes and electromagnetic transition strength are compared with the available experimental data.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

COHERENT BACKSCATTERING VERIFIED NUMERICALLY FOR A FINITE VOLUME OF SPHERICAL PARTICLES

We consider electromagnetic scattering by a spherical volume sparsely and randomly populated by spherical particles of equal size and optical properties. The far-field scattering matrix of the entire volume is computed using an exact method and an approximate method. The former is a direct computer solver of the Maxwell equations called the superposition T-matrix method (STMM). The latter is a solver based on numerical Monte Carlo integration of the ladder and cyclical diagrams appearing in the microphysical theory of radiative transfer and coherent backscattering (RT-CB). The quantitative agreement between the STMM and RT-CB computations provides verification of the RT-CB theory. Prominent backscattering features exhibited by the STMM data cannot be reproduced by keeping only the ladder diagrams of RT. Our results strongly support the CB explanation of opposition brightness and polarization phenomena observed for a class of atmosphereless solar-system objects. Further research is necessary to determine the range of quantitative applicability of the RT-CB theory to densely packed particulate media.

79 ASTRONOMY AND ASTROPHYSICS↗

First-principles study of electron transport in ScN

We investigate the conduction-band structure and electron mobility in rocksalt ScN based on density functional theory. The first-principles band structure allows us to obtain band velocities and effective masses as a function of energy. Electron-phonon scattering is assessed by explicitly computing the q-dependent electron-phonon matrix elements, with the inclusion of the long-range electrostatic interaction. The influence of free-carrier screening on the electron transport is assessed using the random phase approximation. We find a notable enhancement of electron mobility when the carrier concentration exceeds 10 20 cm -3 . We calculate the room-temperature electron mobility in ScN to be 587 cm2 /Vs at low carrier concentrations. When the carrier concentration is increased, the electron mobility starts to decrease significantly around n = 10 19 cm -3 , and drops to 240 cm 2 /Vs at n = 10 21 cm -3 . We also explore the influence of strain in (111)- and (100)-oriented ScN films. For (111) films, we find that a 1.0% compressive epitaxial strain increases the in-plane mobility by 72 cm 2 /Vs and the out-of-plane mobility by 50 cm 2 /Vs. For (100) films, a 1.0% compressive epitaxial strain increases the out-of-plane mobility by as much as 172 cm2 /Vs, but has a weak impact on the in-plane mobility. Our study sheds light on electron transport in ScN at different electron concentrations and shows how strain engineering could increase the electron mobility.

36 MATERIALS SCIENCE↗

Direct Demonstration of the Concept of Unrestricted Effective-Medium Approximation

The modified unrestricted effective-medium refractive index is defined as one that yields accurate values of a representative set of far-field scattering characteristics (including the scattering matrix) for an object made of randomly heterogeneous materials. We validate the concept of the modified unrestricted effective-medium refractive index by comparing numerically exact superposition T-matrix results for a spherical host randomly filled with a large number of identical small inclusions and Lorenz-Mie results for a homogeneous spherical counterpart. A remarkable quantitative agreement between the superposition T-matrix and Lorenz-Mie scattering matrices over the entire range of scattering angles demonstrates unequivocally that the modified unrestricted effective-medium refractive index is a sound (albeit still phenomenological) concept provided that the size parameter of the inclusions is sufficiently small and their number is sufficiently large. Furthermore, it appears that in cases when the concept of the modified unrestricted effective-medium refractive index works, its actual value is close to that predicted by the Maxwell-Garnett mixing rule.

Scattering↗

Identifying Important Ions and Positions in Mass Spectrometry Imaging Data Using CUR Matrix Decompositions

Mass spectrometry imaging enables label-free, high-resolution spatial mapping of the chemical composition of complex, biological samples. Typical experiments require selecting ions and/or positions from the images: ions for fragmentation studies to identify keystone compounds and positions for follow up validation measurements using microdissection or other orthogonal techniques. Unfortunately, with modern imaging machines, these must be selected from an overwhelming amount of raw data. Existing techniques to reduce the volume of data, the most popular of which are principle component analysis and non-negative matrix factorization, have the disadvantage that they return difficult-to-interpret linear combinations of actual data elements. In this work, we show that CX and CUR matrix decompositions can be used directly to address this selection need. CX and CUR matrix decompositions use empirical statistical leverage scores of the input data to provide provably good low-rank approximations of the measured data that are expressed in terms of actual ions and actual positions, as opposed to difficult-to-interpret eigenions and eigenpositions. We show that this leads to effective prioritization of information for both ions and positions. In particular, important ions can be found either by using the leverage scores as a ranking function and using a deterministic greedy selection algorithm or by using the leverage scores as an importance sampling distribution and using a random sampling algorithm; however, selection of important positions from the original matrix performed significantly better when they were chosen with the random sampling algorithm. Also, we show that 20 ions or 40 locations can be used to reconstruct the original matrix to a tolerance of 17% error for a widely studied image of brain lipids; and we provide a scalable implementation of this method that is applicable for analysis of the raw data where there are often more than a million rows and/or columns, which is larger than SVD-based low-rank approximation methods can handle. Lastly, these results introduce the concept of CX/CUR matrix factorizations to mass spectrometry imaging, describing their utility and illustrating principled algorithmic approaches to deal with the overwhelming amount of data generated by modern mass spectrometry imaging.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Multireference Density Functional Theory with Generalized Auxiliary Systems for Ground and Excited States

To describe static correlation, we create a new approach to density functional theory (DFT), which uses a generalized auxiliary system that is of a different symmetry, such as particle number or spin, from that of the physical system. The total energy of the physical system consists of two parts: the energy of the auxiliary system, which is determined with a chosen density functional approximation (DFA), and the excitation energy from an approximate linear response theory that restores the symmetry to that of the physical system, thus rigorously leading to a multideterminant description of the physical system. The electron density of the physical system is different from that of the auxiliary system and is uniquely determined from the functional derivative of the total energy with respect to the external potential. Our energy functional is thus an implicit functional of the physical system density, but an explicit functional of the auxiliary system density. We demonstrate that the total energy minimum and stationary states, describing the ground and excited states of the physical system, can be obtained by a self-consistent optimization with respect to the explicit variable, the generalized Kohn–Sham noninteracting density matrix. We have developed the generalized optimized effective potential method for the self-consistent optimization. Among options of the auxiliary system and the associated linear response theory, reformulated versions of the particle–particle random phase approximation (pp-RPA) and the spin-flip time-dependent density functional theory (SF-TDDFT) are selected for illustration of principle. Numerical findings show that our multireference DFT successfully describes static correlation in bond dissociation and double bond rotation.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

A Low-Rank Multigrid Method for the Stochastic Steady-State Diffusion Problem

We study a multigrid method for solving large linear systems of equations with tensor product structure here. Such systems are obtained from stochastic finite element discretization of stochastic partial differential equations such as the steady-state diffusion problem with random coefficients. When the variance in the problem is not too large, the solution can be well approximated by a low-rank object. In the proposed multigrid algorithm, the matrix iterates are truncated to low rank to reduce memory requirements and computational effort. The technique is proved convergent with an analytic error bound. Numerical experiments show its effectiveness in solving the Galerkin systems compared to the original multigrid solver, especially when the number of degrees of freedom associated with the spatial discretization is large.

97 MATHEMATICS AND COMPUTING↗

Matrix elements in the coupled-cluster approach - With application to low-lying states in Li

A procedure is suggested for evaluating matrix elements of an operator between wavefunctions in the coupled-cluster form. The use of the exponential ansatz leads to compact exponential expressions also for matrix elements. Algorithms are developed for summing all effects of one-particle clusters and certain chains of two-particle clusters (containing the well-known random-phase approximation as a subset). The treatment of one-particle perturbations in single valence states is investigated in detail. As examples the oscillator strength for the 2s-2p transition in Li as well as the hyperfine structure for the two states are studied and compared to earlier work.

Martensson-Pendrill, Ann-Marie↗

A statistical approach to multifield inflation: many-field perturbations beyond slow roll

We study multifield contributions to the scalar power spectrum in an ensemble of six-field inflationary models obtained in string theory. We identify examples in which inflation occurs by chance, near an approximate inflection point, and we compute the primordial perturbations numerically, both exactly and using an array of truncated models. The scalar mass spectrum and the number of fluctuating fields are accurately described by a simple random matrix model. During the approach to the inflection point, bending trajectories and violations of slow roll are commonplace, and 'many-field' effects, in which three or more fields influence the perturbations, are often important. However, in a large fraction of models consistent with constraints on the tilt the signatures of multifield evolution occur on unobservably large scales. Our scenario is a concrete microphysical realization of quasi-single-field inflation, with scalar masses of order H, but the cubic and quartic couplings are typically too small to produce detectable non-Gaussianity. We argue that our results are characteristic of a broader class of models arising from multifield potentials that are natural in the Wilsonian sense.

79 ASTRONOMY AND ASTROPHYSICS↗

Accurate and efficient calculation of excitation energies with the active-space particle-particle random phase approximation

An efficient method for calculating excitation energies based on the particle-particle random phase approximation (ppRPA) is presented. Neglecting the contributions from the high-lying virtual states and the low-lying core states leads to the significantly smaller active-space ppRPA matrix while keeping the error to within 0.05 eV from the corresponding full ppRPA excitation energies. The resulting computational cost is significantly reduced and becomes less than the construction of the non-local Fock exchange potential matrix in the self-consistent-field (SCF) procedure. With only a modest number of active orbitals, the original ppRPA singlet-triplet (ST) gaps as well as the low-lying single and double excitation energies can be accurately reproduced at much reduced computational costs, up to 100 times faster than the iterative Davidson diagonalization of the original full ppRPA matrix. For high-lying Rydberg excitations where the Davidson algorithm fails, the computational savings of active-space ppRPA with respect to the direct diagonalization is even more dramatic. The virtues of the underlying full ppRPA combined with the significantly lower computational cost of the active-space approach will significantly expand the applicability of the ppRPA method to calculate excitation energies at a cost of O(K^{4}), with a prefactor much smaller than a single SCF Hartree-Fock (HF)/hybrid functional calculation, thus opening up new possibilities for the quantum mechanical study of excited state electronic structure of large systems.

74 ATOMIC AND MOLECULAR PHYSICS↗

Application of the target decomposition theorem to a polarimetric random media model

With advances in polarimetric radar measurements of land surfaces, the need for understanding the underlying scattering mechanisms and dominant target features has become the focus of many studies. In particular, the maximum use of the polarimetric information to identify and/or separate parameters related to the surface features such as vegetation thickness, structure, water content, and soil surface characteristics will enhance the possibility of using polarimetric radars for monitoring the earth's surface from space. In this paper, Cloude's decomposition theorem is applied to a polarimetric random media model to simulate the radar measurements of vegetated canopies. The vegetated canopies are modeled as a three layer discrete random medium with leaves and branches in the first layer, tree trunks in the second layer, and a half space of homogeneous ground with rough interface as the bottom layer. The distorted born approximation (DBA) has been used to compute full Mueller matrix of the canopy, using canonical dielectric objects such as thin discs and cylinders as leaves, branches, and trunks, respectively. The Mueller matrix and the derived covariance matrix contain information on the second order statistics of radar signals at various polarizations from the canopy. To decompose the covariance matrix to its constituent targets, the eigenvalues and eigenvectors of the covariance matrix are computed in terms of the physical parameters of the canopy. In addition, each eigenvector explicitly shows the scattering mechanisms such as odd and even reflections in the canopy. Cloude's decomposition theorem is applied using the Pauli spin matrices as a basis and an expression for the degree of disorder or the entropy for the vegetated surface is found. Then, the physical parameters estimated from in situ measurements are used in the random media to obtain realistic covariance matrices. As a result, the sensitivity of the eigenvalue spectrums and the coefficients resulting from the target decomposition theorem to the physical parameters of the canopy are examined and the possible use of Cloude's theorem to estimate vegetation and soil parameters is discussed.

Saatchi, Sasan S.↗