Search NASA⌕ Search

SEARCH · Search NASA

Results for “component analysis”

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 937 records · Page 52

New method for fitting coefficients in standard model effective theory

We present an alternative method for carrying out a principal-component analysis of Wilson coefficients in standard model effective field theory (SMEFT). The method is based on singular-value decomposition (SVD). The SVD method provides information about the sensitivity of experimental observables to physics beyond the standard model that is not accessible in the Fisher-information method. In principle, the SVD method can also have computational advantages over diagonalization of the Fisher information matrix. We demonstrate the SVD method by applying it to the dimension-6 coefficients for the process of top-quark decay to a b quark and a W boson and use this example to illustrate some pitfalls in widely used fitting procedures. We also outline an iterative procedure for applying the SVD method to dimension-8 SMEFT coefficients.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Flow-based likelihoods for non-Gaussian inference

We investigate the use of data-driven likelihoods to bypass a key assumption made in many scientific analyses, which is that the true likelihood of the data is Gaussian. In particular, we suggest using the optimization targets of flow-based generative models, a class of models that can capture complex distributions by transforming a simple base distribution through layers of nonlinearities. We call these flow-based likelihoods (FBL). We analyze the accuracy and precision of the reconstructed likelihoods on mock Gaussian data, and show that simply gauging the quality of samples drawn from the trained model is not a sufficient indicator that the true likelihood has been learned. We nevertheless demonstrate that the likelihood can be reconstructed to a precision equal to that of sampling error due to a finite sample size. We then apply FBLs to mock weak lensing convergence power spectra, a cosmological observable that is significantly non-Gaussian (NG). We find that the FBL captures the NG signatures in the data extremely well, while other commonly used data-driven likelihoods, such as Gaussian mixture models and independent component analysis, fail to do so. This suggests that works that have found small posterior shifts in NG data with data-driven likelihoods such as these could be underestimating the impact of non-Gaussianity in parameter constraints. By introducing a suite of tests that can capture different levels of NG in the data, we show that the success or failure of traditional data-driven likelihoods can be tied back to the structure of the NG in the data. Here, unlike other methods, the flexibility of the FBL makes it successful at tackling different types of NG simultaneously. Because of this, and consequently their likely applicability across datasets and domains, we encourage their use for inference when sufficient mock data are available for training.

79 ASTRONOMY AND ASTROPHYSICS↗

Machine learning BPS spectra and the gap conjecture

We explore statistical properties of Bogomol’nyi-Prasad-Sommerfield q-series for strongly coupled supersymmetric theories that correspond to a particular family of three-manifolds. We discover that gaps between exponents in the -series are statistically more significant at the beginning of the -series compared to gaps that appear in higher powers of. Our observations are obtained by calculating saliencies of -series features used as input data for principal component analysis, which is a standard example of an explainable machine learning technique that allows for a direct calculation and a better analysis of feature saliencies.

97 MATHEMATICS AND COMPUTING↗

High-resolution temporal constraints on the dynamics of dark energy

We use the recent Type Ia supernova, cosmic microwave background and large-scale structure data to shed light on the temporal evolution of the dark energy equation of state w(z) out to redshift one. We constrain the most flexible parametrization of dark energy to date, and include the dark energy perturbations consistently throughout. Interpreting our results via the principal component analysis, we find no significant evidence for dynamical dark energy: the cosmological constant model is consistent with data everywhere between redshift zero and one at 95% C.L.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Measure of the impact of future dark energy experiments based on discriminating power among quintessence models

We evaluate the ability of future data sets to discriminate among different quintessence dark energy models. This approach gives an alternative (and complementary) measure for assessing the impact of future experiments, as compared with the large body of literature that compares experiments in abstract parameter spaces (such as the well-known w{sub 0}-w{sub a} parameters) and more recent work that evaluates the constraining power of experiments on individual parameter spaces of specific quintessence models. We use the Dark Energy Task Force (DETF) models of future data sets and compare the discriminative power of experiments designated by the DETF as stages 2, 3, and 4 (denoting increasing capabilities). Our work reveals a minimal increase in discriminating power when comparing stage 3 to stage 2, but a very striking increase in discriminating power when going to stage 4 (including the possibility of completely eliminating some quintessence models). We also see evidence that even modest improvements over DETF stage 4 (which many believe are realistic) could result in even more dramatic discriminating power among quintessence dark energy models. We develop and demonstrate the technique of using the independently measured modes of the equation of state (derived from principle component analysis) as a common parameter space in which to compare the different quintessence models, and we argue that this technique is a powerful one. We use the PNGB, Exponential, Albrecht-Skordis, and Inverse Tracker (or inverse power law) quintessence models for this work. One of our main results is that the goal of discriminating among these models sets a concrete measure on the capabilities of future dark energy experiments. Experiments have to be somewhat better than DETF stage 4 simulated experiments to fully meet this goal.

79 ASTRONOMY AND ASTROPHYSICS↗

Corrected Article: Measure of the impact of future dark energy experiments based on discriminating power among quintessence models [Phys. Rev. D 78, 043528 (2008)]

We evaluate the ability of future data sets to discriminate among different quintessence dark energy models. This approach gives an alternative (and complementary) measure for assessing the impact of future experiments, as compared with the large body of literature that compares experiments in abstract parameter spaces (such as the well-known w{sub 0}-w{sub a} parameters) and more recent work that evaluates the constraining power of experiments on individual parameter spaces of specific quintessence models. We use the Dark Energy Task Force (DETF) models of future data sets and compare the discriminative power of experiments designated by the DETF as stages 2, 3, and 4 (denoting increasing capabilities). Our work reveals a minimal increase in discriminating power when comparing stage 3 to stage 2, but a very striking increase in discriminating power when going to stage 4 (including the possibility of completely eliminating some quintessence models). We also see evidence that even modest improvements over DETF stage 4 (which many believe are realistic) could result in even more dramatic discriminating power among quintessence dark energy models. We develop and demonstrate the technique of using the independently measured modes of the equation of state (derived from principle component analysis) as a common parameter space in which to compare the different quintessence models, and we argue that this technique is a powerful one. We use the pseudo-Nambu-Goldstone boson, exponential, Albrecht-Skordis, and inverse tracker (or inverse power law) quintessence models for this work. One of our main results is that the goal of discriminating among these models sets a concrete measure on the capabilities of future dark energy experiments. Experiments have to be somewhat better than DETF stage 4 simulated experiments to fully meet this goal.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Probing dark energy dynamics from current and future cosmological observations

We report the constraints on the dark energy equation-of-state w(z) using the latest 'Constitution' SNe sample combined with the WMAP5 and Sloan Digital Sky Survey data. Assuming a flat Universe, and utilizing the localized principal component analysis and the model selection criteria, we find that the {Lambda}CDM model is generally consistent with the current data, yet there exists a weak hint of the possible dynamics of dark energy. In particular, a model predicting w(z)<-1 at z is an element of [0.25,0.5) and w(z)>-1 at z is an element of [0.5,0.75), which means that w(z) crosses -1 in the range of z is an element of [0.25,0.75), is mildly favored at 95% confidence level. Given the best fit model for current data as a fiducial model, we make future forecast from the joint data sets of Joint Dark Energy Mission, Planck, and Large Synoptic Survey Telescope, and we find that the future surveys can reduce the error bars on the w bins by roughly a factor of 10 for a 5-w-bin model.

79 ASTRONOMY AND ASTROPHYSICS↗

How to optimally parametrize deviations from general relativity in the evolution of cosmological perturbations

The next generation of weak lensing surveys will trace the growth of large scale perturbations through a sequence of epochs, offering an opportunity to test general relativity on cosmological scales. We review in detail the parametrization used in MGCAMB to describe the modified growth expected in alternative theories of gravity and generalized dark energy models. We highlight its advantages and examine several theoretical aspects. In particular, we show that the same set of equations can be consistently used on superhorizon and subhorizon linear scales. We also emphasize the sensitivity of data to scale-dependent features in the growth pattern, and propose using principal component analysis to converge on a practical set of parameters which is most likely to detect departures from general relativity. The connection with other parametrizations is also discussed.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Indirect dark matter signatures in the cosmic dark ages. I. Generalizing the bound on s -wave dark matter annihilation from Planck results

Recent measurements of the cosmic microwave background (CMB) anisotropies by Planck provide a sensitive probe of dark matter annihilation during the cosmic dark ages, and specifically constrain the annihilation parameter f eff <σv>=m χ . Using new results (paper II) for the ionization produced by particles injected at arbitrary energies, we calculate and provide f eff values for photons and e + e - pairs injected at keV-TeV energies; the f eff value for any dark matter model can be obtained straightforwardly by weighting these results by the spectrum of annihilation products. This result allows the sensitive and robust constraints on dark matter annihilation presented by the Planck collaboration to be applied to arbitrary dark matter models with s-wave annihilation. Here, we demonstrate the validity of this approach using principal component analysis. As an example, we integrate over the spectrum of annihilation products for a range of Standard Model final states to determine the CMB bounds on these models as a function of dark matter mass, and demonstrate that the new limits generically exclude models proposed to explain the observed high-energy rise in the cosmic ray positron fraction.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Early-Universe constraints on dark matter-baryon scattering and their implications for a global 21 cm signal

Here we present and compare several cosmological constraints on the cross section for elastic scattering between dark matter (DM) and baryons, for cross sections with a range of power-law dependences on the DM-baryon relative velocity v , especially focusing on the case of σ ∝ v - 4 . We study constraints spanning a wide range of epochs in cosmological history, from prerecombination distortions to the blackbody spectrum and anisotropies of the cosmic microwave background (CMB), to modifications to the intergalactic medium temperature and the resulting 21 cm signal, and discuss the allowed signals in the latter channels given the constraints from the former. Furthermore, we improve previous constraints on DM-baryon scattering from the CMB anisotropies, demonstrate via principal component analysis that the effect on the CMB can be written as a simple function of DM mass for v - 4 scattering, and map out the redshifts dominating this signal. We show that given high-redshift constraints on DM-baryon scattering, a v - 4 scaling of the cross section for light DM would be sufficient to explain the deep 21 cm absorption trough recently claimed by the Experiment to Detect the Reionization Step, if 100% of the DM scatters with baryons; we estimate the minimal necessary velocity scaling and find that it is close to v - 4 . For millicharged DM models proposed to explain the observation, where only a small fraction of the DM interacts, we estimate that a PIXIE-like future experiment measuring CMB spectral distortion could test the relevant parameter space.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Reconstruction of the remote dipole and quadrupole fields from the kinetic Sunyaev Zel’dovich and polarized Sunyaev Zel’dovich effects

The kinetic Sunyaev Zel’dovich (kSZ) and polarized Sunyaev Zel’dovich (pSZ) effects are temperature and polarization anisotropies induced by the scattering of cosmic microwave background (CMB) photons from structure in the post-reionization universe. In the case of the kSZ effect, small angular scale anisotropies in the optical depth are modulated by the CMB dipole field, i.e., the CMB dipole observed at each spacetime point, which is sourced by the primordial dipole and especially the local peculiar velocity. In the case of the pSZ effect, similar small-scale anisotropies are modulated by the CMB quadrupole field, which receives contributions from both scalar and tensor modes. Statistical anisotropies in the cross-correlations of CMB temperature and polarization with tracers of the inhomogeneous distribution of electrons provide a means of isolating and reconstructing the dipole and quadrupole fields. In this paper, we present a set of unbiased minimum variance quadratic estimators for the reconstruction of the dipole and quadrupole fields, and forecast the ability of future CMB experiments and large-scale structure surveys to perform this reconstruction. Furthermore, consistent with previous work, we find that a high fidelity reconstruction of the dipole and quadrupole fields over a variety of scales is indeed possible, and demonstrate the sensitivity of the pSZ effect to primordial tensor modes. Using a principle component analysis, we estimate how many independent modes could be accessed in such a reconstruction. We also comment on a few first applications of a detection of the dipole and quadrupole fields, including a reconstruction of the primordial contribution to our locally observed CMB dipole, a test of statistical homogeneity on large scales from the first modes of the quadrupole field, and a reconstruction technique for the primordial potential on the largest scales.

79 ASTRONOMY AND ASTROPHYSICS↗

Symbolic pregression: Discovering physical laws from distorted video

In this work, we present a method for unsupervised learning of equations of motion for objects in raw and optionally distorted unlabeled synthetic video (or, more generally, for discovering and modeling predictable features in time-series data). We first train an autoencoder that maps each video frame into a low-dimensional latent space where the laws of motion are as simple as possible, by minimizing a combination of nonlinearity, acceleration, and prediction error. Differential equations describing the motion are then discovered using Pareto-optimal symbolic regression. We find that our pre-regression (“pregression”) step is able to rediscover Cartesian coordinates of unlabeled moving objects even when the video is distorted by a generalized lens. Using intuition from multidimensional knot theory, we find that the pregression step is facilitated by first adding extra latent space dimensions to avoid topological problems during training and then removing these extra dimensions via principal component analysis. An inertial frame is autodiscovered by minimizing the combined equation complexity for multiple experiments.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Identifying structural changes with unsupervised machine learning methods

In this work, unsupervised machine learning methods are used to identify structural changes using the melting point transition in classical molecular dynamics simulations as an example application of the approach. Dimensionality reduction and clustering methods are applied to instantaneous radial distributions of atomic configurations from classical molecular dynamics simulations of metallic systems over a large temperature range. Principal component analysis is used to dramatically reduce the dimensionality of the feature space across the samples using an orthogonal linear transformation that preserves the statistical variance of the data under the condition that the new feature space is linearly independent. From there, k-means clustering is used to partition the samples into solid and liquid phases through a criterion motivated by the geometry of the reduced feature space of the samples, allowing for an estimation of the melting point transition. This pattern criterion is conceptually similar to how humans interpret the data but with far greater throughput, as the shapes of the radial distributions are different for each phase and easily distinguishable by humans. The transition temperature estimates derived from this machine learning approach produce comparable results to other methods on similarly small system sizes. These results show that machine learning approaches can be applied to structural changes in physical systems.

42 ENGINEERING↗

Large Deviations of the Maximum Eigenvalue for Wishart and Gaussian Random Matrices

We present a Coulomb gas method to calculate analytically the probability of rare events where the maximum eigenvalue of a random matrix is much larger than its typical value. The large deviation function that characterizes this probability is computed explicitly for Wishart and Gaussian ensembles. The method is general and applies to other related problems, e.g., the joint large deviation function for large fluctuations of top eigenvalues. Our results are relevant to widely employed data compression techniques, namely, the principal components analysis. Analytical predictions are verified by extensive numerical simulations.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Random projection using random quantum circuits

The random sampling task performed by Google's Sycamore processor gave us a glimpse of the “quantum supremacy era.” This has definitely shed some light on the power of random quantum circuits in this abstract task of sampling outputs from the (pseudo)random circuits. In this paper, we explore a practical near-term use of local random quantum circuits in dimensional reduction of large low-rank data sets. We make use of the well-studied dimensionality reduction technique called the random projection method. This method has been extensively used in various applications such as image processing, logistic regression, entropy computation of low-rank matrices, etc. We prove that the matrix representations of local random quantum circuits with sufficiently shorter depths [ ∼ O ( n ) ] serve as good candidates for random projection. We demonstrate numerically that their projection abilities are not far off from the computationally expensive classical principal components analysis on MNIST and CIFAR-100 image datasets. We also benchmark the performance of quantum random projection against the commonly used classical random projection in the tasks of dimensionality reduction of image data sets and computing von Neumann entropies of large low-rank density matrices. And finally, using variational quantum singular value decomposition, we demonstrate a near-term implementation of extracting the singular vectors with dominant singular values after quantum random projecting a large low-rank matrix to lower dimensions. All such numerical experiments unequivocally demonstrate the ability of local random circuits to randomize a large Hilbert space at sufficiently shorter depths with robust retention of properties of large data sets in reduced dimensions. Published by the American Physical Society 2024

Kumaran, Keerthi (ORCID:0009000949125721)↗

Active learning approach to simulations of strongly correlated matter with the ghost Gutzwiller approximation

Quantum embedding (QE) methods such as the ghost Gutzwiller approximation (gGA) offer a powerful approach to simulating strongly correlated systems, but come with the computational bottleneck of computing the ground state of an auxiliary embedding Hamiltonian (EH) iteratively. In this work, we introduce an active learning (AL) framework integrated within the gGA to address this challenge. The methodology is applied to the single-band Hubbard model and results in a significant reduction in the number of instances where the EH must be solved. Through a principal component analysis (PCA), we find that the EH parameters form a low-dimensional structure that is largely independent of the geometric specifics of the systems, especially in the strongly correlated regime. Our AL strategy enables us to discover this low-dimensionality structure on the fly, while leveraging it for reducing the computational cost of gGA, laying the groundwork for more efficient simulations of complex strongly correlated materials. Published by the American Physical Society 2024

36 MATERIALS SCIENCE↗

Factors correlating with significant differences between X-ray structures of myoglobin

Conformational differences between myoglobin structures are studied. Most structural differences in whale myoglobin beyond the uncertainty threshold can be correlated with a few specific structural factors. There are always exceptions and a search for additional factors is needed. The results might have serious implications for biological insights from conformational differences. Validation of general ideas about the origins of conformational differences in proteins is critical in order to arrive at meaningful functional insights. Here, principal component analysis (PCA) and distance difference matrices are used to validate some such ideas about the conformational differences between 291 myoglobin structures from sperm whale, horse and pig. Almost all of the horse and pig structures form compact PCA clusters with only minor coordinate differences and outliers that are easily explained. The 222 whale structures form a few dense clusters with multiple outliers. A few whale outliers with a prominent distortion of the GH loop are very similar to the cluster of horse structures, which all have a similar GH-loop distortion apparently owing to intermolecular crystal lattice hydrogen bonds to the GH loop from residues near the distal histidine His64. The variations of the GH-loop coordinates in the whale structures are likely to be owing to the observed alternative intermolecular crystal lattice bond, with the change to the GH loop distorting bonds correlated with the binding of specific ‘unusual’ ligands. Such an alternative intermolecular bond is not observed in horse myoglobins, obliterating any correlation with the ligands. Intermolecular bonds do not usually cause significant coordinate differences and cannot be validated as their universal cause. Most of the native-like whale myoglobin structure outliers can be correlated with a few specific factors. However, these factors do not always lead to coordinate differences beyond the previously determined uncertainty thresholds. The binding of unusual ligands by myoglobin, leading to crystal-induced distortions, suggests that some of the conformational differences between the apo and holo structures might not be ‘functionally important’ but rather artifacts caused by the binding of ‘unusual’ substrate analogs. The causes of P6 symmetry in myoglobin crystals and the relationship between crystal and solution structures are also discussed.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Machine learning inversion from scattering for mechanically driven polymers

A machine learning inversion method is developed for analyzing scattering functions of mechanically driven polymers and extracting the corresponding feature parameters, which include energy parameters and conformation variables. The polymer is modeled as a chain of fixed-length bonds constrained by bending energy, and it is subject to external forces such as stretching and shear. We generate a data set consisting of random combinations of energy parameters, including bending modulus, stretching and shear force, along with Monte Carlo-calculated scattering functions and conformation variables such as end-to-end distance, radius of gyration and off-diagonal component of the gyration tensor. The effects of the energy parameters on the polymer are captured by the scattering function, and principal component analysis ensures the feasibility of the machine learning inversion. Finally, we train a Gaussian process regressor using part of the data set as a training set and validate the trained regressor for inversion using the rest of the data. The regressor successfully extracts the feature parameters.

Gaussian process regressors↗