Search NASA⌕ Search

SEARCH · Search NASA

Results for “computational statistics”

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 739 records · Page 41

LyMAS: Predicting large-scale Lyα forest statistics from the dark matter density field

We describe Lyα Mass Association Scheme (LyMAS), a method of predicting clustering statistics in the Lyα forest on large scales from moderate-resolution simulations of the dark matter (DM) distribution, with calibration from high-resolution hydrodynamic simulations of smaller volumes. We use the 'Horizon-MareNostrum' simulation, a 50 h {sup –1} Mpc comoving volume evolved with the adaptive mesh hydrodynamic code RAMSES, to compute the conditional probability distribution P(F{sub s} |δ {sub s}) of the transmitted flux F{sub s} , smoothed (one-dimensionally, 1D) over the spectral resolution scale, on the DM density contrast δ {sub s}, smoothed (three-dimensionally, 3D) over a similar scale. In this study we adopt the spectral resolution of the SDSS-III Baryon Oscillation Spectroscopic Survey (BOSS) at z = 2.5, and we find optimal results for a DM smoothing length σ = 0.3 h {sup –1} Mpc (comoving). In its simplest form, LyMAS draws randomly from the hydro-calibrated P(F{sub s} |δ {sub s}) to convert DM skewers into Lyα forest pseudo-spectra, which are then used to compute cross-sightline flux statistics. In extended form, LyMAS exactly reproduces both the 1D power spectrum and one-point flux distribution of the hydro simulation spectra. Applied to the MareNostrum DM field, LyMAS accurately predicts the two-point conditional flux distribution and flux correlation function of the full hydro simulation for transverse sightline separations as small as 1 h {sup –1} Mpc, including redshift-space distortion effects. It is substantially more accurate than a deterministic density-flux mapping ({sup F}luctuating Gunn-Peterson Approximation{sup )}, often used for large-volume simulations of the forest. With the MareNostrum calibration, we apply LyMAS to 1024{sup 3} N-body simulations of a 300 h {sup –1} Mpc and 1.0 h {sup –1} Gpc cube to produce large, publicly available catalogs of mock BOSS spectra that probe a large comoving volume. LyMAS will be a powerful tool for interpreting 3D Lyα forest data, thereby transforming measurements from BOSS and other massive quasar absorption surveys into constraints on dark energy, DM, space geometry, and intergalactic medium physics.

79 ASTRONOMY AND ASTROPHYSICS↗

Topological Cacti: Visualizing Contour-based Statistics

Contours, the connected components of level sets, play an important role in understanding the global structure of a scalar field. In particular their nestingbehavior and topology-often represented in form of a contour tree-have been used extensively for visualization and analysis. However, traditional contour trees onlyencode structural properties like number of contours or the nesting of contours, but little quantitative information such as volume or other statistics. Here we use thesegmentation implied by a contour tree to compute a large number of per-contour (interval) based statistics of both the function defining the contour tree as well asother co-located functions. We introduce a new visual metaphor for contour trees, called topological cacti, that extends the traditional toporrery display of acontour tree to display additional quantitative information as width of the cactus trunk and length of its spikes. We apply the new technique to scalar fields ofvarying dimension and different measures to demonstrate the effectiveness of the approach.

97 MATHEMATICS AND COMPUTING↗

Ultrafast processing of pixel detector data with machine learning frameworks

Modern photon science performed at high repetition rate free-electron laser (FEL) facilities and beyond relies on 2D pixel detectors operating at increasing frequencies (towards 100 kHz at LCLS-II) and producing rapidly increasing amounts of data (towards TB/s). This data must be rapidly stored for offline analysis and summarized in real time for online feedback to the scientists. While at LCLS all raw data has been stored, at LCLS-II this would lead to a prohibitive cost; instead, enabling real time processing of pixel detector data (dark, gain, common mode, background, charge summing, subpixel position, photon counting, data summarization) allows reducing the size and cost of online processing, offline processing and storage by orders of magnitude while preserving full photon information. This could be achieved by taking advantage of the compressibility of sparse data typical for LCLS-II applications. Faced with a similar big data challenge a decade ago, computer vision stimulated revolutionary advances in machine learning hardware and software. We investigated if these developments are useful in the field of data processing for high speed pixel detectors and found that typical deep learning models and autoencoder architectures failed to yield useful noise reduction while preserving full photon information, presumably because of the very different statistics and feature sets in computer vision and radiation imaging. However, the raw performance of modern frameworks like Tensorflow inspired us to redesign in Tensorflow mathematically equivalent versions of the state-of-the-art, “classical” algorithms used at LCLS. The novel Tensorflow models resulted in elegant, compact and hardware agnostic code, gaining 1 to 2 orders of magnitude faster processing on an inexpensive consumer GPU, reducing by 3 orders of magnitude the projected cost of online analysis and compression without photon loss at LCLS-II. The novel Tensorflow models also enabled ongoing development of a pipelined hardware system expected to yield an additional 3 to 4 orders of magnitude speedup, necessary for meeting the data acquisition and storage requirements at LCLS-II, potentially enabling acquiring every single FEL pulse at full speed. Computer vision a decade ago was dominated by hand-crafted filters; their structure inspired the deep learning revolution resulting in modern deep convolutional networks; similarly, our novel Tensorflow filters provide inspiration for designing future deep learning models for ultrafast and efficient processing and classification of pixel detector images at FEL facilities.

47 OTHER INSTRUMENTATION↗

TECA: Petascale pattern recognition for climate science

Climate Change is one of the most pressing challenges facing humanity in the 21st century. Climate simulations provide us with a unique opportunity to examine effects of anthropogenic emissions. Highresolution climate simulations produce “Big Data”: contemporary climate archives are ≈ 5PB in size and we expect future archives to measure on the order of Exa-Bytes. In this work, we present the successful application of TECA (Toolkit for Extreme Climate Analysis) framework, for extracting extreme weather patterns such as Tropical Cyclones, Atmospheric Rivers and Extra-Tropical Cyclones from TB-sized simulation datasets. TECA has been run at full-scale on Cray XE6 and IBM BG/Q systems, and has reduced the runtime for pattern detection tasks from years to hours. TECA has been utilized to evaluate the performance of various computational models in reproducing the statistics of extreme weather events, and for characterizing the change in frequency of storm systems in the future.

54 ENVIRONMENTAL SCIENCES↗

Examination of two methods for statistical analysis of data with magnitude and direction emphasizing vestibular research applications

When the dependent (or response) variable response variable in an experiment has direction and magnitude, one approach that has been used for statistical analysis involves splitting magnitude and direction and applying univariate statistical techniques to the components. However, such treatment of quantities with direction and magnitude is not justifiable mathematically and can lead to incorrect conclusions about relationships among variables and, as a result, to flawed interpretations. This note discusses a problem with that practice and recommends mathematically correct procedures to be used with dependent variables that have direction and magnitude for 1) computation of mean values, 2) statistical contrasts of and confidence intervals for means, and 3) correlation methods.

Vestibule↗

Optimized blind gamma-ray pulsar searches at fixed computing budget

The sensitivity of blind gamma-ray pulsar searches in multiple years worth of photon data, as from the Fermi LAT, is primarily limited by the finite computational resources available. Addressing this 'needle in a haystack' problem, here we present methods for optimizing blind searches to achieve the highest sensitivity at fixed computing cost. For both coherent and semicoherent methods, we consider their statistical properties and study their search sensitivity under computational constraints. The results validate a multistage strategy, where the first stage scans the entire parameter space using an efficient semicoherent method and promising candidates are then refined through a fully coherent analysis. We also find that for the first stage of a blind search incoherent harmonic summing of powers is not worthwhile at fixed computing cost for typical gamma-ray pulsars. Further enhancing sensitivity, we present efficiency-improved interpolation techniques for the semicoherent search stage. Via realistic simulations we demonstrate that overall these optimizations can significantly lower the minimum detectable pulsed fraction by almost 50% at the same computational expense.

79 ASTRONOMY AND ASTROPHYSICS↗

Techniques for generation of control and guidance signals derived from optical fields, part 2

The development is reported of a high resolution technique for the detection and identification of landmarks from spacecraft optical fields. By making use of nonlinear regression analysis, a method is presented whereby a sequence of synthetic images produced by a digital computer can be automatically adjusted to provide a least squares approximation to a real image. The convergence of the method is demonstrated by means of a computer simulation for both elliptical and rectangular patterns. Statistical simulation studies with elliptical and rectangular patterns show that the computational techniques developed are able to at least match human pattern recognition capabilities, even in the presence of large amounts of noise. Unlike most pattern recognition techniques, this ability is unaffected by arbitrary pattern rotation, translation, and scale change. Further development of the basic approach may eventually allow a spacecraft or robot vehicle to be provided with an ability to very accurately determine its spatial relationship to arbitrary known objects within its optical field of view.

Hemami, H.↗

Efficient computation of N -point correlation functions in D dimensions

We present efficient algorithms for computing the N-point correlation functions (NPCFs) of random fields in arbitrary D-dimensional homogeneous and isotropic spaces. Such statistics appear throughout the physical sciences and provide a natural tool to describe stochastic processes. Typically, algorithms for computing the NPCF components have $\mathscr O$(n N ) complexity (for a dataset containing n particles); their application is thus computationally infeasible unless N is small. By projecting the statistic onto a suitably defined angular basis, we show that the estimators can be written in a separable form, with complexity $\mathscr O$(n 2 ) or $\mathscr O$(n g log n g ) if evaluated using a Fast Fourier Transform on a grid of size n g . Our decomposition is built upon the D-dimensional hyperspherical harmonics; these form a complete basis on the (D – 1) sphere and are intrinsically related to angular momentum operators. Concatenation of (N – 1) such harmonics gives states of definite combined angular momentum, forming a natural separable basis for the NPCF. As N and D grow, the number of basis components quickly becomes large, providing a practical limitation to this (and all other) approaches: However, the dimensionality is greatly reduced in the presence of symmetries; for example, isotropic correlation functions require only states of zero combined angular momentum. We provide a Julia package implementing our estimators and show how they can be applied to a variety of scenarios within cosmology and fluid dynamics. The efficiency of such estimators will allow higher-order correlators to become a standard tool in the analysis of random fields.

97 MATHEMATICS AND COMPUTING↗

Potential energy surface fitting by a statistically localized, permutationally invariant, local interpolating moving least squares method for the many-body potential: Method and application to N{sub 4}

Fitting potential energy surfaces to analytic forms is an important first step for efficient molecular dynamics simulations. Here, we present an improved version of the local interpolating moving least squares method (L-IMLS) for such fitting. Our method has three key improvements. First, pairwise interactions are modeled separately from many-body interactions. Second, permutational invariance is incorporated in the basis functions, using permutationally invariant polynomials in Morse variables, and in the weight functions. Third, computational cost is reduced by statistical localization, in which we statistically correlate the cutoff radius with data point density. We motivate our discussion in this paper with a review of global and local least-squares-based fitting methods in one dimension. Then, we develop our method in six dimensions, and we note that it allows the analytic evaluation of gradients, a feature that is important for molecular dynamics. The approach, which we call statistically localized, permutationally invariant, local interpolating moving least squares fitting of the many-body potential (SL-PI-L-IMLS-MP, or, more simply, L-IMLS-G2), is used to fit a potential energy surface to an electronic structure dataset for N{sub 4}. We discuss its performance on the dataset and give directions for further research, including applications to trajectory calculations.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

ON COMPUTING UPPER LIMITS TO SOURCE INTENSITIES

A common problem in astrophysics is determining how bright a source could be and still not be detected in an observation. Despite the simplicity with which the problem can be stated, the solution involves complicated statistical issues that require careful analysis. In contrast to the more familiar confidence bound, this concept has never been formally analyzed, leading to a great variety of often ad hoc solutions. Here we formulate and describe the problem in a self-consistent manner. Detection significance is usually defined by the acceptable proportion of false positives (background fluctuations that are claimed as detections, or Type I error), and we invoke the complementary concept of false negatives (real sources that go undetected, or Type II error), based on the statistical power of a test, to compute an upper limit to the detectable source intensity. To determine the minimum intensity that a source must have for it to be detected, we first define a detection threshold and then compute the probabilities of detecting sources of various intensities at the given threshold. The intensity that corresponds to the specified Type II error probability defines that minimum intensity and is identified as the upper limit. Thus, an upper limit is a characteristic of the detection procedure rather than the strength of any particular source. It should not be confused with confidence intervals or other estimates of source intensity. This is particularly important given the large number of catalogs that are being generated from increasingly sensitive surveys. We discuss, with examples, the differences between these upper limits and confidence bounds. Both measures are useful quantities that should be reported in order to extract the most science from catalogs, though they answer different statistical questions: an upper bound describes an inference range on the source intensity, while an upper limit calibrates the detection process. We provide a recipe for computing upper limits that applies to all detection algorithms.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Improved Pyrotechnic Shock Scaling Method Based on Shock Response Spectrum and Statistical Energy Analysis

It has been more than forty years since the pyrotechnic shock scaling method was introduced. The scaling method estimates the attenuation of the Shock Response Spectrum (SRS) based on the distance from the source, structural configurations, types of structural joints and interfaces, and intervening structure [1]. The method has been successfully used in the spacecraft community countless times and still is frequently used to develop pyrotechnic shock requirements at various levels of assembly of a spacecraft. However, since the method was derived empirically from a limited set of shock test data [2], the aerospace community has been looking for an alternative approach. In this paper, a computational scaling method based on Statistical Energy Analysis (SEA) is re-visited. SEA is traditionally used as a method for investigating the diffusion of acoustic and vibratory energies of a system at steady-state [3]. Previously, an approach using the SRS as an acceleration constraint condition to the SEA model was introduced to estimate the attenuation from a shock source [4,5]. In the current investigation, additional examples are provided to further validate the approach. In addition, responses in the time domain are produced by the Local Modal Phase Reconstruction (LMPR) approach which was recently introduced to the community [6].

Fernandez, Juan P.↗

Veridical data science

Building and expanding on principles of statistics, machine learning, and scientific inquiry, we propose the predictability, computability, and stability (PCS) framework for veridical data science. Our framework, composed of both a workflow and documentation, aims to provide responsible, reliable, reproducible, and transparent results across the data science life cycle. The PCS workflow uses predictability as a reality check and considers the importance of computation in data collection/storage and algorithm design. It augments predictability and computability with an overarching stability principle. Stability expands on statistical uncertainty considerations to assess how human judgment calls impact data results through data and model/algorithm perturbations. As part of the PCS workflow, we develop PCS inference procedures, namely PCS perturbation intervals and PCS hypothesis testing, to investigate the stability of data results relative to problem formulation, data cleaning, modeling decisions, and interpretations. We illustrate PCS inference through neuroscience and genomics projects of our own and others. Moreover, we demonstrate its favorable performance over existing methods in terms of receiver operating characteristic (ROC) curves in high-dimensional, sparse linear model simulations, including a wide range of misspecified models. Finally, we propose PCS documentation based on R Markdown or Jupyter Notebook, with publicly available, reproducible codes and narratives to back up human choices made throughout an analysis. The PCS workflow and documentation are demonstrated in a genomics case study available on Zenodo.

97 MATHEMATICS AND COMPUTING↗

Subspace Iteration Method for Complex Eigenvalue Problems with Nonsymmetric Matrices in Aeroelastic System

Modern airplane design is a multidisciplinary task which combines several disciplines such as structures, aerodynamics, flight controls, and sometimes heat transfer. Historically, analytical and experimental investigations concerning the interaction of the elastic airframe with aerodynamic and in retia loads have been conducted during the design phase to determine the existence of aeroelastic instabilities, so called flutter .With the advent and increased usage of flight control systems, there is also a likelihood of instabilities caused by the interaction of the flight control system and the aeroelastic response of the airplane, known as aeroservoelastic instabilities. An in -house code MPASES (Ref. 1), modified from PASES (Ref. 2), is a general purpose digital computer program for the analysis of the closed-loop stability problem. This program used subroutines given in the International Mathematical and Statistical Library (IMSL) (Ref. 3) to compute all of the real and/or complex conjugate pairs of eigenvalues of the Hessenberg matrix. For high fidelity configuration, these aeroelastic system matrices are large and compute all eigenvalues will be time consuming. A subspace iteration method (Ref. 4) for complex eigenvalues problems with nonsymmetric matrices has been formulated and incorporated into the modified program for aeroservoelastic stability (MPASES code). Subspace iteration method only solve for the lowest p eigenvalues and corresponding eigenvectors for aeroelastic and aeroservoelastic analysis. In general, the selection of p is ranging from 10 for wing flutter analysis to 50 for an entire aircraft flutter analysis. The application of this newly incorporated code is an experiment known as the Aerostructures Test Wing (ATW) which was designed by the National Aeronautic and Space Administration (NASA) Dryden Flight Research Center, Edwards, California to research aeroelastic instabilities. Specifically, this experiment was used to study an instability known as flutter. ATW was a small-scale airplane wing comprised of an airfoil and wing tip boom. This wing was formulated based on a NACA-65A004 airfoil shape with a 3.28 aspect ratio. The wing had a span of 18 inch with root chord length of 13.2 inch and tip chord length of 8.7 inch. The total area of this wing was 197 square inch. The wing tip boom was a 1 inch diameter hollow tube of length 21.5 inch. The total weight of the wing was 2.66 lbs.

Pak, Chan-gi↗

Discovery of Activities via Statistical Clustering of Fixation Patterns

Human behavior often consists of a series of distinct activities, each characterized by a unique pattern of interaction with the visual environment. This is true even in a restricted domain, such as a pilot flying an airplane; in this case, activities with distinct visual signatures might be things like communicating, navigating, monitoring, etc. We propose a novel analysis method for gaze-tracking data, to perform blind discovery of these hypothetical activities. We compare, not individual fixations, but groups of fixations aggregated over a fixed time interval (Tau). We assume that the environment has been divided into a finite set of discrete areas-of-interest (AOIs). For a given time interval, we compute the proportion of time spent fixating each AOI, resulting in an N-dimensional vector, where N is the number of AOIs. These proportions can be converted to integer counts by multiplying by Tau divided by the average fixation duration, a parameter that we fix at 283 milliseconds. We compare different intervals by computing the chi-squared statistic. The p-value associated with the statistic is the likelihood of observing the data under the hypothesis that the data in the two intervals were generated by a single process with a single set of probabilities governing the fixation of each AOI. We cluster the intervals, first by merging adjacent intervals that are sufficiently similar, optionally shifting the boundary between non-merged intervals to maximize the difference. Then we compare and cluster non-adjacent intervals. The method is evaluated using synthetic data generated by a hand-crafted set of activities. While the method generally finds more activities than put into the simulation, we have obtained agreement as high as 80 percent between the inferred activity labels and ground truth.

Eye Movements↗

General Diffuse-Interface Theories and an Approach to Predictive Tumor Growth Modeling

While a large and growing literature exists on mathematical and computational models of tumor growth, to date tumor growth models are largely qualitative in nature, and fall far short of being able to provide predictive results important in life-and-death decisions. This is largely due to the enormous complexity of evolving biological and chemical processes in living tissue and the complex interactions of many cellular and vascular constituents in living organisms. Several new technologies have emerged, however, which could lead to significant progress in this important area: (i) the development of so-called phase-field, or diffuse-interface models, which can be developed using continuum mixture theory, and which provide a general framework for modeling the action of multiple interacting constituents. These are based on generalizations of the Cahn–Hilliard models for spinodal decomposition, and have been used recently in certain tumor growth theories; (ii) the emergence of predictive computational methods based on the use of statistical methods for calibration, model validation, and uncertainty quantification; (iii) advances in imaging, experimental cell biology, and other medical observational methodologies; and (iv) the advent of petascale computing that makes possible the resolution of features at scales and at speeds that were unattainable only a short time in the past. Here we develop a general phenomenological thermomechanical theory of mixtures that employs phase-field or diffuse interface models of surface energies and reactions and which provides a framework for generalizing existing theories of the types that are in use in tumor growth modeling. In principle, the framework provides for the effects of M solid constituents, which may undergo large deformations, and for the effect of N - M fluid constituents, which could include highly nonlinear, non-Newtonian fluids. We then describe several special cases which have the potential of providing acceptable models of tumor growth. We then describe the beginning steps of the development of Bayesian methods for statistical calibration, model validation, and uncertainty quantification, which, with further work, could produce a truly predictive tool for studying tumor growth. In particular, we outline the processes of statistical calibration and validation that can be employed to determine if tumor growth models, drawn from the broad class of models developed here, are valid for prediction of key quantities of interest critical to making decisions on various medical protocols. We also describe how uncertainties in such key quantities of interest can be quantified in ways that can be used to establish confidence in predicted outcomes.

59 BASIC BIOLOGICAL SCIENCES↗