Search NASA⌕ Search

SEARCH · Search NASA

Results for “random sampling”

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 1,387 records · Page 77

Modeling the Swift BAT trigger algorithm with machine learning

To draw inferences about gamma-ray burst (GRB) source populations based on Swift observations, it is essential to understand the detection efficiency of the Swift burst alert telescope (BAT). This study considers the problem of modeling the Swift/BAT triggering algorithm for long GRBs, a computationally expensive procedure, and models it using machine learning algorithms. A large sample of simulated GRBs from Lien et al. is used to train various models: random forests, boosted decision trees (with AdaBoost), support vector machines, and artificial neural networks. The best models have accuracies of ≳97% (≲3% error), which is a significant improvement on a cut in GRB flux, which has an accuracy of 89.6% (10.4% error). These models are then used to measure the detection efficiency of Swift as a function of redshift z, which is used to perform Bayesian parameter estimation on the GRB rate distribution. We find a local GRB rate density of n{sub 0}∼0.48{sub −0.23}{sup +0.41} Gpc{sup −3} yr{sup −1} with power-law indices of n{sub 1}∼1.7{sub −0.5}{sup +0.6} and n{sub 2}∼−5.9{sub −0.1}{sup +5.7} for GRBs above and below a break point of z{sub 1}∼6.8{sub −3.2}{sup +2.8}. This methodology is able to improve upon earlier studies by more accurately modeling Swift detection and using this for fully Bayesian model fitting.

79 ASTRONOMY AND ASTROPHYSICS↗

3FGL demographics outside the galactic plane using supervised machine learning: pulsar and dark matter subhalo interpretations

Nearly one-third of the sources listed in the Third Fermi Large Area Telescope (LAT) catalog (3FGL) remain unassociated. It is possible that predicted and even unanticipated gamma-ray source classes are present in these data waiting to be discovered. Taking advantage of the excellent spectral capabilities achieved by the Fermi LAT, we use machine-learning classifiers (Random Forest and XGBoost) to pinpoint potentially novel source classes in the unassociated 3FGL sample outside the Galactic plane. Here we report a total of 34 high-confidence Galactic candidates at |b|⩾5{sup ∘}. The currently favored standard astrophysical interpretations for these objects are pulsars or low-luminosity globular clusters hosting millisecond pulsars (MSPs). Yet these objects could also be interpreted as dark matter annihilation taking place in ultra-faint dwarf galaxies or dark matter subhalos. Unfortunately, Fermi LAT spectra are not sufficient to break degeneracies between the different scenarios. Careful visual inspection of archival optical images reveals no obvious evidence for low-luminosity globular clusters or ultra-faint dwarf galaxies inside the 95% error ellipses. If these are pulsars, this would bring the total number of MSPs at |b|⩾5{sup ∘} to 106, down to an energy flux ≈4.0 × 10{sup −12} erg cm{sup −2} s{sup −1} between 100 MeV and 100 GeV. We find this number to be in excellent agreement with predictions from a new population synthesis of MSPs that predicts 100–126 high-latitude 3FGL MSPs depending on the choice of high-energy emission model. If, however, these are dark matter substructures, we can place upper limits on the number of Galactic subhalos surviving today and on dark matter annihilation cross sections. These limits are beginning to approach the canonical thermal relic cross section for dark matter particle masses below ∼100 GeV in the bottom quark (b b-bar ) annihilation channel.

79 ASTRONOMY AND ASTROPHYSICS↗

Plasma electrolytes, pH, and ECG during and after exhaustive exercise.

Ten men worked on a bicycle ergometer at increasing work loads to exhaustion in 15 min. Each performed one test breathing air and another with added CO2 in random sequence. ECG was recorded during exercise and for 30 min of recovery. Arterial samples for blood gases, pH, and electrolytes were drawn at rest, in the last minute of exercise and at 1, 4, 10, 20, and 30 min thereafter. A striking increase in the amplitude of T and P waves was observed reaching a maximum in the first 2 min after exercise. All electrolytes measured were increased at the end of exercise, most markedly potassium (60%) and phosphorus (53%). Potassium dropped faster than all others to below resting values in 4 min coinciding with the lowest levels in plasma bicarbonate. ECG alterations were not closely related in time with any single factor such as potassium, but appeared to reflect an interaction of the transient mineral and acid-base imbalance during and immediately following exhaustive exercise.

Coester, N.↗

Radar error statistics for the space shuttle

Radar error statistics of C-band and S-band that are recommended for use with the groundtracking programs to process space shuttle tracking data are presented. The statistics are divided into two parts: bias error statistics, using the subscript B, and high frequency error statistics, using the subscript q. Bias errors may be slowly varying to constant. High frequency random errors (noise) are rapidly varying and may or may not be correlated from sample to sample. Bias errors were mainly due to hardware defects and to errors in correction for atmospheric refraction effects. High frequency noise was mainly due to hardware and due to atmospheric scintillation. Three types of atmospheric scintillation were identified: horizontal, vertical, and line of sight. This was the first time that horizontal and line of sight scintillations were identified.

Lear, W. M.↗

Some methods of estimating a coverage parameter

To demonstrate the high reliability of fault-tolerant computing systems, experimental testing can be performed by injecting faults into their hardware and observing the times needed to detect and isolate failed components and reconfigure the good components. Coverage, a parameter measuring continued system success, is the probability that recovery occurs before other, perhaps catastrophic, component failures occur. Methods of estimating coverage are given for the case of randomly selecting a subset of the pins or chips for testing. Pin-level samples of times to recovery are treated by the Type I censoring model often used in life testing.

Lee, L. D.↗

Simulation of deep one- and two-dimensional redshift surveys

It is shown that slice or pencil-beam redshift surveys of galaxies can be simulated in a box with nonequal sides. This method saves a lot of computer time and memory while providing essentially the same results as from whole-cube simulations. A 2457.6/h Mpc-long rod (out to a redshift z = 0.58 in two opposite directions) is simulated using the standard biased cold dark matter model as an example to mimic the recent deep pencil-beam surveys by Broadhurst et al. (1990). The structures (spikes) seen in these simulated samples occur when the narrow pencil-beam pierces walls, filaments, and clusters appearing randomly along the line-of-sight. A statistical test for goodness of fit to a periodic lattice has been applied to the observations and the simulations. It is found that the statistical significance level (P = 15.4 percent) is not strong enough to reject the null hypothesis that the observations and the simulations were drawn at random from the same set.

Park, Changbom↗

An investigation of error correcting techniques for OMV and AXAF

The original objectives of this project were to build a test system for the NASA 255/223 Reed/Solomon encoding/decoding chip set and circuit board. This test system was then to be interfaced with a convolutional system at MSFC to examine the performance of the concantinated codes. After considerable work, it was discovered that the convolutional system could not function as needed. This report documents the design, construction, and testing of the test apparatus for the R/S chip set. The approach taken was to verify the error correcting behavior of the chip set by injecting known error patterns onto data and observing the results. Error sequences were generated using pseudo-random number generator programs, with Poisson time distribution between errors and Gaussian burst lengths. Sample means, variances, and number of un-correctable errors were calculated for each data set before testing.

Ingels, Frank↗

Probabilistic nonlinear finite element analysis of composite structures

A probabilistic finite element analysis procedure for laminated composite shells is developed. A total Lagrangian finite element formulation, employing a degenerated three-dimensional laminated composite shell element with the full Green-Lagrange strains and first-order shear deformable kinematics, is used. The first-order second-moment technique for probabilistic finite element analysis of random fields is employed, and results are presented in the form of mean and variance of the structural response. Reliability calculations are made by using the first-order reliability method combined with sensitivity derivatives from the finite element analysis. Both ply-level and micromechanics-level random variables are incorporated, the latter by means of the Aboudi micromechanics model. Two sample problems are solved to verify the accuracy of the procedures developed and to quantify the variability of certain material type/structure combinations. In general, the procedure is quite effective in determining the response statistics and reliability for linear and geometric nonlinear behavior of laminated composite shells.

Engelstad, S. P.↗

Modeling the Swift Bat Trigger Algorithm with Machine Learning

To draw inferences about gamma-ray burst (GRB) source populations based on Swift observations, it is essential to understand the detection efficiency of the Swift burst alert telescope (BAT). This study considers the problem of modeling the Swift / BAT triggering algorithm for long GRBs, a computationally expensive procedure, and models it using machine learning algorithms. A large sample of simulated GRBs from Lien et al. is used to train various models: random forests, boosted decision trees (with AdaBoost), support vector machines, and artificial neural networks. The best models have accuracies of greater than or equal to 97 percent (less than or equal to 3 percent error), which is a significant improvement on a cut in GRB flux, which has an accuracy of 89.6 percent (10.4 percent error). These models are then used to measure the detection efficiency of Swift as a function of redshift z, which is used to perform Bayesian parameter estimation on the GRB rate distribution. We find a local GRB rate density of n (sub 0) approaching 0.48 (sup plus 0.41) (sub minus 0.23) per cubic gigaparsecs per year with power-law indices of n (sub 1) approaching 1.7 (sup plus 0.6) (sub minus 0.5) and n (sub 2) approaching minus 5.9 (sup plus 5.7) (sub minus 0.1) for GRBs above and below a break point of z (redshift) (sub 1) approaching 6.8 (sup plus 2.8) (sub minus 3.2). This methodology is able to improve upon earlier studies by more accurately modeling Swift detection and using this for fully Bayesian model fitting.

gamma-ray burst: general – gamma-rays: general â↗

Mapping National Forest Aboveground Biomass in Mexico by Integrating GEDI, Sentinel‐1 and Sentinel‐2 Data

Accurate mapping of forest aboveground biomass density (AGBD) is required to better understand the role of forests in the global carbon cycle and to support international policies for climate change mitigation and adaptation. Mexico is one of the countries having great potential for the United Nations Programme on Reducing Emissions from Deforestation and Forest Degradation (or UN-REDD program) and there is a growing demand for unbiased Monitoring Reporting Verification systems at a national level. As an effort under NASA’s Carbon Monitoring System (CMS) program, we developed a machine learning model using multi-stream remote sensing measurements as well as topographic data to create a high spatial resolution AGBD map (~100 m) over Mexico (circa 2020). The remote sensing data includes Global Ecosystem Dynamic Investigation (GEDI) lidar, Sentinel 1 Synthetic-Aperture Radar (SAR), and Sentinel-2 multispectral imagery (MSI). GEDI onboard the International Space Station provides unprecedented forest structure and AGBD sampling datasets for model training and validation practices. Our analysis indicates that the developed random forest model can capture 63 % of the spatial variation (RMSE = 33.7 Mg/ha) of AGBD of Mexican forests. We find that shortwave infrared bands of Sentinel-2 MSI and topographical variables from elevation data are the most important variables in the developed AGBD model. Our study highlights methodological opportunities in synergistic uses of multiple sensors for large-scale forest AGBD mapping and shows potential for retrospective analysis and operational monitoring of forest AGBD and its dynamics.

Taejin Park↗

No Galaxy Left Behind: Accurate Measurements with the Faintest Objects in the Dark Energy Survey

Accurate statistical measurement with large imaging surveys has traditionally required throwing away a sizable fraction of the data. This is because most measurements have relied on selecting nearly complete samples, where variations in the composition of the galaxy population with seeing, depth, or other survey characteristics are small. Here, we introduce a new measurement method that aims to minimize this wastage, allowing precision measurement for any class of detectable stars or galaxies. Moreover, our proposal was implemented in BALROG, software which embeds fake objects in real imaging to accurately characterize measurement biases. We demonstrate this technique with an angular clustering measurement using Dark Energy Survey (DES) data. We first show that recovery of our injected galaxies depends on a variety of survey characteristics in the same way as the real data. We then construct a flux-limited sample of the faintest galaxies in DES, chosen specifically for their sensitivity to depth and seeing variations. Using the synthetic galaxies as randoms in the Landy–Szalay estimator suppresses the effects of variable survey selection by at least two orders of magnitude. With this correction, our measured angular clustering is found to be in excellent agreement with that of a matched sample from much deeper, higher resolution space-based Cosmological Evolution Survey (COSMOS) imaging; over angular scales of 0$_.^°$004 < θ < 0$_.^°$2, we find a best-fitting scaling amplitude between the DES and COSMOS measurements of 1.00 ± 0.09. We expect this methodology to be broadly useful for extending measurements’ statistical reach in a variety of upcoming imaging surveys.

79 ASTRONOMY AND ASTROPHYSICS↗

Enabling probabilistic learning on manifolds through double diffusion maps

Here, we present a generative learning framework for probabilistic sampling that extends Probabilistic Learning on Manifolds (PLoM), which is designed to generate statistically consistent realizations of a random vector in a finite-dimensional Euclidean space, informed by a (representative) set of observations. In its original form, PLoM constructs a reduced-order probabilistic model by combining three main components: (a) kernel density estimation to approximate the underlying probability measure, (b) Diffusion Maps to characterize the manifold of the data, and (c) a reduced-order Itô Stochastic Differential Equation (ISDE) to sample from the learned distribution. However, its sampling dynamics are posed in the ambient space and the retained number of reduced coordinates is chosen by projection-reconstruction error. In practice, this often (i) requires more coordinates than the data’s intrinsic dimension to achieve stable sampling and (ii) lacks a smooth, basis-independent lifting back to the data domain; moreover, standard Diffusion Maps emphasize harmonic eigenfunctions and can miss non-harmonic latent structure. We address these limitations by decoupling geometry learning from sampling: a first Diffusion Maps pass identifies non-harmonic coordinates on which we formulate a full-order ISDE directly in the latent space, while Double Diffusion Maps captures multiscale geometric features and Geometric Harmonics (GH) learns a smooth lifting map to the ambient variables that is independent of the particular diffusion basis. This hybrid design preserves the system’s dynamical richness with a compact geometric representation and enables principled out-of-sample inference. The effectiveness and robustness of the proposed method are illustrated through two numerical studies: one based on data generated from two-dimensional Hermite polynomial functions and another based on high-fidelity simulations of a detonation wave in a reactive flow.

Double diffusion maps↗

Stochastic modeling and statistical calibration with model error and scarce data

This paper introduces a procedure to assess the predictive accuracy of stochastic models subject to model error and sparse data. Model error is introduced as uncertainty on the coefficients of appropriate polynomial chaos expansions (PCE). The error associated with finite sample size allows us to conceive of these coefficients as statistics of the data that we describe as random variables whose influence on output quantities of interest is evaluated through the extended polynomial chaos expansion (EPCE). A Bayesian data assimilation scheme is introduced to update these expansions by considering the resulting nested chaos expansion as a hierarchical probabilistic model. Stochastic models of quantities of interest (QoI) are thus constructed and efficiently evaluated. Here, the Metropolis–Hastings Markov chain Monte Carlo procedure is used to sample the posterior. Two illustrative analytical and numerical problems are used to demonstrate the proposed approach.

Bayesian inference↗

Biological monitoring of smoke exposure among wildland firefighters: A pilot study comparing urinary methoxyphenols with personal exposures to carbon monoxide, particulate matter, and levoglucosan.

Urinary methoxyphenols (MPs) have been proposed as biomarkers of woodsmoke exposure. However, few field studies have been undertaken to evaluate the relationship between woodsmoke exposure and urinary MP concentrations. We conducted a pilot study at the US Forest ServiceFSavannah River Site, in which carbon monoxide (CO), levoglucosan (LG), and particulate matter (PM2.5) exposures were measured in wildland firefighters on prescribedburn days. Pre- and post-shift urine samples were collected from each subject, and cross-shift changes in creatinine-corrected urinary MP concentrations were calculated. Correlations between exposure measures and creatine-adjusted urinary MP concentrations were explored, and regression models were developed relating changes in urinary MP concentrations to measured exposure levels. Full-shift measurements were made on 13 firefighters over 20 work shifts in winter 2004 at the US Forest Service Savannah River site, a National Environmental Research Park. The average workshift length across the 20 measured shifts was 701±95 min. LG and CO exposures were significantly correlated for samples where the filter measurement captured at least 60% of the work shift (16 samples), as well as for the smaller set of full-shift exposure samples (n¼9). PM2.5 and CO exposures were not significantly correlated, and LG and PM2.5 exposures were only significantly correlated for samples representing at least 60% of the work shift. Creatinine-corrected urinary concentrations for 20 of the 22 MPs showed cross-shift increases, with 14 of these changes showing statistical significance. Individual and summed creatinine-adjusted guaiacol urinary MPs were highly associated with CO (and, to a lesser degree, LG) exposure levels, and random-effects regression models including CO and LG exposure levels explained up to 80% of the variance in cross-shift changes in summed creatinine-adjusted guaiacol urinary MP concentrations. Although limited by the small sample size, this pilot study demonstrates that urinary MP concentrations may be effective biomarkers of occupational exposure to wood smoke among wildland firefighters.

60 APPLIED LIFE SCIENCES↗

Randomized Algorithms for Scientific Computing (RASC)

Randomized algorithms have propelled advances in artificial intelligence (AI) and represent a foundational research area in advancing AI for Science. Future advancements in DOE Office of Science priority areas such as climate science, astrophysics, fusion, advanced materials, combustion, and quantum computing all require randomized algorithms for surmounting challenges of complexity, robustness, and scalability. Advances in data collection and numerical simulation have changed the dynamics of scientific research and motivate the need for randomized algorithms. For instance, advances in imaging technologies such as X-ray ptychography, electron microscopy, electron energy loss spectroscopy, or adaptive optics lattice light-sheet microscopy collect hyperspectral imaging and scattering data in terabytes, at breakneck speed enabled by state-of-the-art detectors. The data collection is exceptionally fast compared with its analysis. Likewise, advances in high-performance architectures have made exascale computing a reality and changed the economies of scientific computing in the process. Floating-point operations that create data are essentially free in comparison with data movement. Thus far, most approaches have focused on creating faster hardware. Ironically, this faster hardware has exacerbated the problem by making data still easier to create. Under such an onslaught, scientists often resort to heuristic deterministic sampling schemes (e.g., low-precision arithmetic, sampling every nth element) and sacrifice potentially valuable accuracy. Dramatically better results can be achieved via randomized algorithms, reducing the data size as much as or more than naive deterministic subsampling can achieve, while retaining the high accuracy of computing on the full data set. By randomized algorithms we mean those algorithms that employ some form of randomness in internal algorithmic decisions to accelerate time to solution, increase scalability, or improve reliability. Examples include matrix sketching for solving large-scale least-squares problems (see Figure 1) and stochastic gradient descent for training machine learning models. We are not recommending heuristic methods but rather randomized algorithms that have certificates of correctness and probabilistic guarantees of optimality and near-optimality. Such approaches can be useful beyond acceleration, for example, in understanding how to avoid measure zero worst-case scenarios that plague methods such as QR matrix factorization.

97 MATHEMATICS AND COMPUTING↗

Intensity of sample processing methods impacts wastewater SARS-CoV-2 whole genome amplicon sequencing outcomes

Wastewater SARS-CoV-2 surveillance has been deployed since the beginning of the COVID-19 pandemic to monitor the dynamics in virus burden in local communities. Genomic surveillance of SARS-CoV-2 in wastewater, particularly efforts aimed at whole genome sequencing for variant tracking and identification, are still challenging due to low target concentration, complex microbial and chemical background, and lack of robust nucleic acid recovery experimental procedures. The intrinsic sample limitations are inherent to wastewater and are thus unavoidable. Here, we use a statistical approach that couples correlation analyses to a random forest-based machine learning algorithm to evaluate potentially important factors associated with wastewater SARS-CoV-2 whole genome amplicon sequencing outcomes, with a specific focus on the breadth of genome coverage. We collected 182 composite and grab wastewater samples from the Chicago area between November 2020 to October 2021. Samples were processed using a mixture of processing methods reflecting different homogenization intensities (HA + Zymo beads, HA + glass beads, and Nanotrap), and were sequenced using one of the two library preparation kits (the Illumina COVIDseq kit and the QIAseq DIRECT kit). Technical factors evaluated using statistical and machine learning approaches include sample types, certain sample intrinsic features, and processing and sequencing methods. The results suggested that sample processing methods could be a predominant factor affecting sequencing outcomes, and library preparation kits was considered a minor factor. Finally, a synthetic SARS-CoV-2 RNA spike-in experiment was performed to validate the impact from processing methods and suggested that the intensity of the processing methods could lead to different RNA fragmentation

60 APPLIED LIFE SCIENCES↗

The topology of large-scale structure. V - Two-dimensional topology of sky maps

A 2D algorithm is applied to observed sky maps and numerical simulations. It is found that when topology is studied on smoothing scales larger than the correlation length, the topology is approximately in agreement with the random phase formula for the 2D genus-threshold density relation, G2(nu) varies as nu(e) exp-nu-squared/2. Some samples show small 'meatball shifts' similar to those seen in corresponding 3D observational samples and similar to those produced by biasing in cold dark matter simulations. The observational results are thus consistent with the standard model in which the structure in the universe today has grown from small fluctuations caused by random quantum noise in the early universe.

Gott, J. R., III↗