Search NASA⌕ Search

SEARCH · Search NASA

Results for “Tensor factorizations”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

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

At least 73 records · Page 4
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Tensor Network Quantum Virtual Machine for Simulating Quantum Circuits at Exascale

The numerical simulation of quantum circuits is an indispensable tool for development, verification, and validation of hybrid quantum-classical algorithms intended for near-term quantum co-processors. The emergence of exascale high-performance computing (HPC) platforms presents new opportunities for pushing the boundaries of quantum circuit simulation. Here, we present a modernized version of the Tensor Network Quantum Virtual Machine (TNQVM) that serves as the quantum circuit simulation backend in the eXtreme-scale ACCelerator (XACC) framework. The new version is based on the scalable tensor network processing library ExaTN (Exascale Tensor Networks). It provides multiple configurable quantum circuit simulators that perform either an exact quantum circuit simulation via the full tensor network contraction or an approximate simulation via a suitably chosen tensor factorization scheme. Upon necessity, stochastic noise modeling from real quantum processors is incorporated into the simulations by modeling quantum channels with Kraus tensors. By combining the portable XACC quantum programming frontend and the scalable ExaTN numerical processing backend, we introduce an end-to-end virtual quantum development environment that can scale from laptops to future exascale platforms. We report initial benchmarks of our framework, which include a demonstration of the distributed execution, incorporation of quantum decoherence models, and simulation of the random quantum circuits used for the certification of quantum supremacy on Google’s Sycamore superconducting architecture.

Nguyen, Thien↗
DOE OSTI2025

A Latent-Variable Formulation of the Poisson Canonical Polyadic Tensor Model: Maximum Likelihood Estimation and Fisher Information

We establish parameter inference for the Poisson canonical polyadic (PCP) tensor model through a latent-variable formulation. Our approach exploits the observation that any random PCP tensor can be derived by marginalizing an unobservable random tensor of one dimension larger. The loglikelihood of this larger dimensional tensor, referred to as the “complete” loglikelihood, is comprised of multiple rank one PCP loglikelihoods. Using this methodology, we first derive maximum likelihood estimators for the PCP model and demonstrate that several existing algorithms for fitting non-negative matrix and tensor factorizations are Expectation-Maximization algorithms. Next, we derive the observed and expected Fisher information matrices for the PCP model. The Fisher information provides us crucial insights into the well-posedness of the tensor model, such as the role that tensor rank plays in identifiability and indeterminacy. For the special case of rank one PCP models, we demonstrate that these results are greatly simplified.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2015

Further investigation of g factors for the lead monofluoride ground state

We report the results of our theoretical study and analysis of earlier experimental data for the g-factor tensor components of the ground 2 II 1/2 state of the free PbF radical. These values obtained both within the relativistic coupled-cluster method combined with the generalized relativistic effective core potential approach and with our fit of the experimental data from [R. J. Mawhorter, B. S. Murphy, A. L. Baum, T. J. Sears, T. Yang, P. M. Rupasinghe, C. P. McRaven, N. E. Shafer-Ray, L. D. Alphei, and J.-U. Grabow, Phys. Rev. A 84, 022508 (2011); A. L. Baum, B.A. thesis, Pomona College, 2011]. The obtained results agree very well with each other but contradict the previous fit performed in the cited works. Our final prediction for g factors is G ∥ =0.081(5),G ⊥ =–0.27(1).

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2016

Wormholes, emergent gauge fields, and the weak gravity conjecture

This paper revisits the question of reconstructing bulk gauge fields as boundary operators in AdS/CFT. In the presence of the ormhole dual to the thermo field double state of two CFTs, the existence of bulk gauge fields is in some tension with the microscopic tensor factorization of the Hilbert space. Here, I explain how this tension can be resolved by splitting the gauge field into charged constituents, and I argue that this leads to a new argument for the "principle of completeness", which states that the charge lattice of a gauge theory coupled to gravity must be fully populated. I also claim that it leads to a new motivation for (and a clarification of) the "weak gravity conjecture", which I interpret as a strengthening of this principle. This setup gives a simple example of a situation where describing low-energy bulk physics in CFT language requires knowledge of high-energy bulk physics. Furthermore, this contradicts to some extent the notion of "effective conformal field theory", but in fact is an expected feature of the resolution of the black hole information problem. An analogous factorization issue exists also for the gravitational field, and I comment on several of its implications for reconstructing black hole interiors and the emergence of spacetime more generally.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2015

Hilbert space structure in quantum gravity: an algebraic perspective

If quantum gravity respects the principles of quantum mechanics, suitably generalized, it may be that a more viable approach to the theory is through identifying the relevant quantum structures rather than by quantizing classical spacetime. Here, this viewpoint is supported by difficulties of such quantization, and by the apparent lack of a fundamental role for locality. In finite or discrete quantum systems, important structure is provided by tensor factorizations of the Hilbert space. However, even in local quantum field theory properties of the generic type III von Neumann algebras and of long range gauge fields indicate that factorization of the Hilbert space is problematic. Instead it is better to focus on the structure of the algebra of observables, and in particular on its subalgebras corresponding to regions. This paper suggests that study of analogous algebraic structure in gravity gives an important perspective on the nature of the quantum theory. Significant departures from the subalgebra structure of local quantum field theory are found, working in the correspondence limit of long-distances/low-energies. Particularly, there are obstacles to identifying commuting algebras of localized operators. In addition to suggesting important properties of the algebraic structure, this and related observations pose challenges to proposals of a fundamental role for entanglement.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2019

Quantum-First Gravity

This paper elaborates on an intrinsically quantum approach to gravity, which begins with a general framework for quantum mechanics and then seeks to identify additional mathematical structure on Hilbert space that is responsible for gravity and other phenomena. A key principle in this approach is that of correspondence: this structure should reproduce spacetime, general relativity, and quantum field theory in a limit of weak gravitational fields. A central question is that of "Einstein separability," and asks how to define mutually independent subsystems, e.g. through localization. Familiar definitions involving tensor products or operator subalgebras do not clearly accomplish this in gravity, as is seen in the correspondence limit. Instead, gravitational behavior, particularly gauge invariance, suggests a network of Hilbert subspaces related via inclusion maps, contrasting with other approaches based on tensor-factorized Hilbert spaces. Finally, any such localization structure is also expected to place strong constraints on evolution, which are also supplemented by the constraint of unitarity.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2022

Accelerating high-order mesh optimization using finite element partial assembly on GPUs

In this paper we present a new GPU-oriented mesh optimization method based on high order finite elements. Our approach relies on node movement with fixed topology, through the Target-Matrix Optimization Paradigm (TMOP) and uses a global nonlinear solve over the whole computational mesh, i.e., all mesh nodes are moved together. A key property of the method is that the mesh optimization process is recast in terms of finite element operations, which allows us to utilize recent advances in the field of GPU-accelerated high order finite element algorithms. For example, we reduce data motion by using tensor factorization and matrix-free methods, which have superior performance characteristics compared to traditional full finite element matrix assembly and offer advantages for GPU based HPC hardware. Furthermore, we describe the major mathematical components of the method along with their efficient GPU-oriented implementation. In addition, we propose an easily reproducible mesh optimization test that can serve as a performance benchmark for the mesh optimization community.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Robust Tensor Hypercontraction of the Particle–Particle Ladder Term in Equation-of-Motion Coupled Cluster Theory

One method of representing a high-rank tensor as a (hyper-)product of lower-rank tensors is the tensor hypercontraction (THC) method of Hohenstein et al. This strategy has been found to be useful for reducing the polynomial scaling of coupled-cluster methods by representation of a four-dimensional tensor of electron-repulsion integrals in terms of five two-dimensional matrices. Pierce et al. have already shown that the application of a robust form of THC to the particle–particle ladder (PPL) term reduces the cost of this term in couple-cluster singles and doubles (CCSD) from O(N 6 ) to O(N 5 ) with negligible errors in energy with respect to the density-fitted variant. In this work, we have implemented the least-squares variant of THC (LS-THC) which does not require a nonlinear tensor factorization, including the robust form (R-LS-THC), for the calculation of the excitation and electron attachment energies using equation-of-motion coupled cluster methods EOMEE-CCSD and EOMEA-CCSD, respectively. We have benchmarked the effect of the R-LS-THC-PPL approximation on excitation energies using the comprehensive QUEST database and the accuracy of electron attachment energies using the NAB22 database. Here, we find that errors on the order of 1 meV are achievable with a reduction in total calculation time of approximately 5x.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2023

Symmetry-projected cluster mean-field theory applied to spin systems

We introduce S z spin-projection based on cluster mean-field theory and apply it to the ground state of strongly correlated spin systems. In cluster mean-fields, the ground state wavefunction is written as a factorized tensor product of optimized cluster states. In previous work, we have focused on unrestricted cluster mean-field, where each cluster is S z symmetry adapted. We here remove this restriction by introducing a generalized cluster mean-field (GcMF) theory, where each cluster is allowed to access all S z sectors, breaking S z symmetry. In addition, a projection scheme is used to restore global S z , which gives rise to the S z spin-projected generalized cluster mean-field (S z GcMF). Both of these extensions contribute to accounting for inter-cluster correlations. We benchmark these methods on the 1D, quasi-2D, and 2D J 1 – J 2 and XXZ Heisenberg models. Furthermore, our results indicate that the new methods (GcMF and S z GcMF) provide a qualitative and semi-quantitative description of the Heisenberg lattices in the regimes considered, suggesting them as useful references for further inter-cluster correlations, which are discussed in this work.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2019

Machine Learning

Presentation overview: Why Machine Learning?; Why ML-Informed Physics?; Why Physics-Informed ML?; LANL ML strategy related to National & Energy Security; ML-Informed Physics for climate modeling and induced seismicity (LDRD-DR); Physics-Informed ML for blind source separation (LDRD-DR); ML forecasting production from tight rock reservoirs (Chevron)

58 GEOSCIENCES↗
DOE OSTI2011

B to tensor meson form factors in the perturbative QCD approach

We calculate the B{sub u,d,s}{yields}T form factors within the framework of the perturbative QCD approach, where T denotes a light tensor meson with J{sup P}=2{sup +}. Because of the similarities between the wave functions of a vector and a tensor meson, the factorization formulas of B{yields}T form factors can be obtained from the B{yields}V transition through a replacement rule. As a consequence, we find that these two sets of form factors have the same signs and correlated q{sup 2}-dependence behaviors. At q{sup 2}=0 point, the B{yields}T form factors are smaller than the B{yields}V ones, in accordance with the experimental data of radiative B decays. In addition, we use our results for the form factors to explore semilteptonic B{yields}Tl{nu}{sub l} decays and the branching fractions can reach the order 10{sup -4}.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2024

Geometry-aware training of factorized layers in tensor Tucker format

Reducing parameter redundancies in neural network architectures is crucial for achieving feasible computational and memory requirements during train and inference of large networks. Given its easy implementation and flexibility, one promising approach is layer factorization, which reshapes weight tensors into a matrix format and parameterizes it as the product of two rank-r matrices. However, this family of approaches often requires an initial full-model warm-up phase, prior knowledge of a feasible rank, and it is sensitive to parameter initialization.In this work, we introduce a novel approach to train the factors of a Tucker decomposition of the weight tensors. Our training proposal proves to be optimal in locally approximating the original unfactorized dynamics and stable for the initialization. Furthermore, the rank of each mode is dynamically updated during training.We provide a theoretical analysis of the algorithm, showing convergence, approximation and local descent guarantees. The method's performance is further illustrated through a variety of experiments, showing remarkable training compression rates and comparable or even better performance than the full baseline and alternative layer factorization strategies.

Zangrando, Emanuele [Gran Sasso Science Institute ↗
DOE OSTI2010

Tensor charges and form factors of SU(3) baryons in the self-consistent SU(3) chiral quark-soliton model

We investigate the tensor form factors of the baryon octet within the framework of the chiral quark-soliton model, emphasizing those of the nucleon, taking linear 1/N{sub c} rotational as well as linear m{sub s} corrections into account, and applying the symmetry-conserving quantization. We explicitly calculate the tensor form factors H{sub T}{sup q}(Q{sup 2}) corresponding to the generalized parton distributions H{sub T}(x,{xi},t). The tensor form factors are obtained for the momentum transfer up to Q{sup 2{<=}}1 GeV{sup 2} and at a renormalization scale of 0.36 GeV{sup 2}. We find for the tensor charges {delta}u=1.08, {delta}d=-0.32, and {delta}s=-0.01 and discuss their physical consequences, comparing them with those from other models. Results for tensor charges for the baryon octet are also given.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2020

Generalized Canonical Polyadic Tensor Decomposition

Tensor decomposition is a fundamental unsupervised machine learning method in data science, with applications including network analysis and sensor data processing. This work develops a generalized canonical polyadic (GCP) low-rank tensor decomposition that allows other loss functions besides squared error. For instance, we can use logistic loss or Kullback--Leibler divergence, enabling tensor decomposition for binary or count data. We present a variety of statistically motivated loss functions for various scenarios. We provide a generalized framework for computing gradients and handling missing data that enables the use of standard optimization methods for fitting the model. Furthermore, we demonstrate the flexibility of the GCP decomposition on several real-world examples including interactions in a social network, neural activity in a mouse, and monthly rainfall measurements in India.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2010

Anomalous tensor magnetic moments and form factors of the proton in the self-consistent chiral quark-soliton model

We investigate the form factors of the chiral-odd nucleon matrix element of the tensor current. In particular, we aim at the anomalous tensor magnetic form factors of the nucleon within the framework of the SU(3) and SU(2) chiral quark-soliton model. We consider 1/N{sub c} rotational corrections and linear effects of SU(3) symmetry breaking with the symmetry-conserving quantization employed. We first obtain the results of the anomalous tensor magnetic moments for the up and down quarks: {kappa}{sub T}{sup u}=3.56 and {kappa}{sub T}{sup d}=1.83, respectively. The strange anomalous tensor magnetic moment is yielded to be {kappa}{sub T}{sup s}=0.2{approx}-0.2, that is compatible with zero. We also calculate the corresponding form factors {kappa}{sub T}{sup q}(Q{sup 2}) up to a momentum transfer Q{sup 2{<=}}1 GeV{sup 2} at a renormalization scale of 0.36 GeV{sup 2}.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2020

Theoretical and Phenomenological Studies of the Nucleon Structure in High Energy Reactions

The understanding of the internal structure of the proton and other strongly interacting particles is at the forefront of modern nuclear physics research. Generalized Parton Distribution Functions (GPDs) are a powerful tool to advance the understanding of hadron structure. In addition to the information about the one- dimensional collinear momentum distributions of partons (quarks, anti-quarks, and gluons) known from studies of high energy deep-inelastic reactions, GPDs also carry information on the distribution of partons in the transverse plane, and allow us in this way to access the three-dimensional structure of the nucleon. GPDs can be studied in hard-exclusive reactions and contain also information on the energy-momentum tensor form factors which will allow us to gain insights on quantities like pressure or angular momentum distribution inside the nucleon. The goal of this thesis is to deepen our understanding of the three-dimensional structure of the nucleon. We investigate energy-momentum tensor form factors and densities, and all leading-twist GPDs in the bag model. This quark model provides a consistent theoretical framework to investigate many general concepts that have recently attracted interest, and allows one to study insightful limits like the large-Nc limit, heavy-quark limit or the non-relativistic limit. Another important aspect of this thesis is the discussion of the monopole and quadrupole contributions to the angular momentum density. Finally, the description of pseudoscalar meson production in exclusive processes in terms of GPDs is discussed in the Goloskokov-Kroll model and implemented in the PARTONS framework, a software development project which will provide direct support for experiments at the Jefferson National Lab and the future Electron-Ion Collider.

Tezgin, Kemal↗
DOE OSTI2011

'Forbidden' reflections in resonant diffraction of synchrotron radiation in yttrium aluminum garnet Y{sub 3}Al{sub 5}O{sub 12}

The purely resonant Bragg reflections (13, 13, 0) and (14, 0, 0) in yttrium aluminum garnet Y{sub 3}Al{sub 5}O{sub 12} at energies near the K absorption edge of yttrium have been studied experimentally and theoretically. The anisotropic tensor atomic factor of yttrium corresponding to dipole-dipole resonance transitions depends on three independent parameters changing with energy. The intensities of the reflections (14, 0, 0) and (13, 13, 0) are shown to depend on the parameter difference f{sub 1}(E) - f{sub 2}(E) and the parameter f{sub 3}(E), respectively, which are attributable to distortions of the wave functions of the excited atoms and change greatly with photon energy E. Studying various reflections has allowed one to determine the various components of the tensor atomic factor and to compare them with the results of numerical calculations.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗