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 559 records · Page 31

Characterization of Electron Microscopes with Binary Pseudo-random Multilayer Test Samples

Verification of the reliability of metrology data from high quality X-ray optics requires that adequate methods for test and calibration of the instruments be developed. For such verification for optical surface profilometers in the spatial frequency domain, a modulation transfer function (MTF) calibration method based on binary pseudo-random (BPR) gratings and arrays has been suggested [1] and [2] and proven to be an effective calibration method for a number of interferometric microscopes, a phase shifting Fizeau interferometer, and a scatterometer [5]. Here we describe the details of development of binary pseudo-random multilayer (BPRML) test samples suitable for characterization of scanning (SEM) and transmission (TEM) electron microscopes. We discuss the results of TEM measurements with the BPRML test samples fabricated from a WiSi2/Si multilayer coating with pseudo-randomly distributed layers. In particular, we demonstrate that significant information about the metrological reliability of the TEM measurements can be extracted even when the fundamental frequency of the BPRML sample is smaller than the Nyquist frequency of the measurements. The measurements demonstrate a number of problems related to the interpretation of the SEM and TEM data. Note that similar BPRML test samples can be used to characterize X-ray microscopes. Corresponding work with X-ray microscopes is in progress.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Characterization of electron microscopes with binary pseudo-random multilayer test samples

Verification of the reliability of metrology data from high quality X-ray optics requires that adequate methods for test and calibration of the instruments be developed. For such verification for optical surface profilometers in the spatial frequency domain, a modulation transfer function (MTF) calibration method based on binarypseudo-random (BPR) gratings and arrays has been suggested and and proven to be an effective calibration method for a number of interferometric microscopes, a phase shifting Fizeau interferometer, and a scatterometer. Here we describe the details of development of binarypseudo-random multilayer (BPRML) test samples suitable for characterization of scanning (SEM) and transmission (TEM) electron microscopes. We discuss the results of TEM measurements with the BPRML test samples fabricated from a WiSi{sub 2}/Si multilayer coating with pseudo-randomly distributed layers. In particular, we demonstrate that significant information about the metrological reliability of the TEM measurements can be extracted even when the fundamental frequency of the BPRML sample is smaller than the Nyquist frequency of the measurements. The measurements demonstrate a number of problems related to the interpretation of the SEM and TEM data. Note that similar BPRML testsamples can be used to characterize X-ray microscopes. Corresponding work with X-ray microscopes is in progress.

36 MATERIALS SCIENCE↗

Transcriptomics-based Machine Learning (ML) Analysis Predicts Space-Exposed Murine Livers

Limited sample sizes, high data dimensionality, and sensitivity to technical and biological variability of next generation sequencing (NGS), typically limits machine learning (ML) approaches in spaceflight studies that include radiation effects. However, pooling smaller studies while addressing intra- and inter-study variabilities allows for ML predictive modeling. Here, integration methods were applied to whole transcriptome shotgun sequencing (RNA-seq) data from six mouse liver GeneLab datasets (GLDS) (n ranging from 6 to 39 samples) from with a total of 81 spaceflight and ground-control samples to determine top features (i.e. genes) relevant to spaceflight including the effect of radiation exposure. RNASeq counts were normalized for each study, then merged and scaled across all datasets. Data dimensionality was reduced using a minimum redundancy maximum relevance (MRMR) methodology. Redundancy and relevance were computed using the Pearson correlation and F-statistic, respectively. The top 100 MRMR features were used to predict spaceflight vs. ground-control samples using Random Forest (RF), Support Vector Machine (SVM), and Linear Discriminant Analysis (LDA) classifiers with 5-fold cross validation (CV). Principal component analysis (PCA) on the complete feature set versus the MRMR features shows separation between spaceflight samples and ground controls (Figure 1A). The ML-based gene sets were compared against differential gene expression results obtained with DESeq2 from individual GLDS. Using all features or randomly sampled subsets at matching set sizes with MRMR, a maximum classifier accuracy of 69% on the test set over 5 folds. For all classifiers, CV training using at least the top 30 MRMR genes show minimum 89% accuracy and 0.95 AUC value on the test set over 5 folds (Figure 1B). Baseline set analysis on differentially expressed genes (DEGs) identified using padj ≤ 0.05 show 295 DEGs that overlap at least two studies and 13 DEGs that overlap three studies (Figure 1C). Set analysis between the top 100 MRMR features and the DEGs showed 47 genes that overlap at least one study and 24 genes that overlap two studies. Over-representation analysis showed overlapping biological processes related to fatty acid and lipid metabolism which may indicate these processes in the response to spaceflight stressors. MRMR feature selection for the selected ML methods improve performance relative to a classifier built on all features or randomly sampled subsets. Permutation feature importance within the decorrelated MRMR features showed concordance in feature ranking between ML methods. A challenge of applying ML methods across heterogeneous NGS data is accounting for signal:noise. Here, signal validation across studies was shown by intersecting sets between top MRMR genes and DEGs from DESeq2 analysis. Non-intersecting sets introduce opportunity to explore genes relevant to differentiating space flight exposed groups and implementing ML methods across existing NGS datasets may overcome sample size limitations.

Machine Learning↗

Transcriptomics-based Machine Learning Analysis Predicts Space-Exposed Murine Livers

Limited sample sizes, high data dimensionality, and sensitivity to technical and biological variability of next generation sequencing (NGS), typically limits machine learning (ML) approaches in spaceflight studies that include radiation effects. However, pooling smaller studies while addressing intra- and inter-study variabilities allows for ML predictive modeling. Here, integration methods were applied to whole transcriptome shotgun sequencing (RNA-seq) data from six mouse liver GeneLab datasets (GLDS) (n ranging from 6 to 39 samples) from with a total of 81 spaceflight and ground-control samples to determine top features (i.e. genes) relevant to spaceflight including the effect of radiation exposure. RNASeq counts were normalized for each study, then merged and scaled across all datasets. Data dimensionality was reduced using a minimum redundancy maximum relevance (MRMR) methodology. Redundancy and relevance were computed using the Pearson correlation and F-statistic, respectively. The top 100 MRMR features were used to predict spaceflight vs. ground-control samples using Random Forest (RF), Support Vector Machine (SVM), and Linear Discriminant Analysis (LDA) classifiers with 5-fold cross validation (CV). Principal component analysis (PCA) on the complete feature set versus the MRMR features shows separation between spaceflight samples and ground controls (Figure 1A). The ML-based gene sets were compared against differential gene expression results obtained with DESeq2 from individual GLDS. Using all features or randomly sampled subsets at matching set sizes with MRMR, a maximum classifier accuracy of 69% was shown on the test set over 5 folds. For all classifiers, CV training using at least the top 30 MRMR genes show minimum 89% accuracy and 0.95 AUC value on the test set over 5 folds (Figure 1B). Baseline set analysis on differentially expressed genes (DEGs) identified using padj ≤ 0.05 show 295 DEGs that overlap at least two studies and 13 DEGs that overlap three studies (Figure 1C). Set analysis between the top 100 MRMR features and the DEGs showed 47 genes that overlap at least one study and 24 genes that overlap two studies. Over-representation analysis showed overlapping biological processes related to fatty acid and lipid metabolism which may indicate these processes in the response to spaceflight stressors. MRMR feature selection for the selected ML methods improve performance relative to a classifier built on all features or randomly sampled subsets. Permutation feature importance within the decorrelated MRMR features showed concordance in feature ranking between ML methods. A challenge of applying ML methods across heterogeneous NGS data is accounting for signal:noise. Here, signal validation across studies was shown by intersecting sets between top MRMR genes and DEGs from DESeq2 analysis. Non-intersecting sets introduce opportunity to explore genes relevant to differentiating space flight exposed groups and implementing ML methods across existing NGS datasets may overcome sample size limitations.

Machine Learning↗

Transcriptomics-based Machine Learning Analysis Predicts Space-Exposed Murine Livers

Limited sample sizes, high data dimensionality, and sensitivity to technical and biological variability of next generation sequencing (NGS), typically limits machine learning (ML) approaches in spaceflight studies that include radiation effects. However, pooling smaller studies while addressing intra- and inter-study variabilities allows for ML predictive modeling. Here, integration methods were applied to whole transcriptome shotgun sequencing (RNA-seq) data from six mouse liver GeneLab datasets (GLDS) (n ranging from 6 to 39 samples) from with a total of 81 spaceflight and ground-control samples to determine top features (i.e. genes) relevant to spaceflight including the effect of radiation exposure. RNASeq counts were normalized for each study, then merged and scaled across all datasets. Data dimensionality was reduced using a minimum redundancy maximum relevance (MRMR) methodology. Redundancy and relevance were computed using the Pearson correlation and F-statistic, respectively. The top 100 MRMR features were used to predict spaceflight vs. ground-control samples using Random Forest (RF), Support Vector Machine (SVM), and Linear Discriminant Analysis (LDA) classifiers with 5-fold cross validation (CV). Principal component analysis (PCA) on the complete feature set versus the MRMR features shows separation between spaceflight samples and ground controls (Figure 1A). The ML-based gene sets were compared against differential gene expression results obtained with DESeq2 from individual GLDS. Using all features or randomly sampled subsets at matching set sizes with MRMR, a maximum classifier accuracy of 69% was shown on the test set over 5 folds. For all classifiers, CV training using at least the top 30 MRMR genes show minimum 89% accuracy and 0.95 AUC value on the test set over 5 folds (Figure 1B). Baseline set analysis on differentially expressed genes (DEGs) identified using padj ≤ 0.05 show 295 DEGs that overlap at least two studies and 13 DEGs that overlap three studies (Figure 1C). Set analysis between the top 100 MRMR features and the DEGs showed 47 genes that overlap at least one study and 24 genes that overlap two studies. Over-representation analysis showed overlapping biological processes related to fatty acid and lipid metabolism which may indicate these processes in the response to spaceflight stressors. MRMR feature selection for the selected ML methods improve performance relative to a classifier built on all features or randomly sampled subsets. Permutation feature importance within the decorrelated MRMR features showed concordance in feature ranking between ML methods. A challenge of applying ML methods across heterogeneous NGS data is accounting for signal:noise. Here, signal validation across studies was shown by intersecting sets between top MRMR genes and DEGs from DESeq2 analysis. Non-intersecting sets introduce opportunity to explore genes relevant to differentiating space flight exposed groups and implementing ML methods across existing NGS datasets may overcome sample size limitations.

Machine Learning↗

Assessing Risk Due to Small Sample Size in Probability of Detection Analysis Using Tolerance Intervals

Small sample size (e.g.6-30) poses risk in results of probability of detection (POD) analysis using tolerance intervals. This method is also called as the limited sample or LS POD. The analysis is performed either during NDE procedure qualification or for assessment of reliability of an NDE procedure. The risk is primarily due to sampling error. Smaller samples are not likely to be random to the population or representative of the population. The small samples are likely to be biased. Biased samples have smaller standard deviation compared to the population. POD analysis with small biased sample can lead to overestimation of POD. Many sampling schemes are available in statistics to mitigate sampling risk. Primary objective of POD analysis is to determine a decision threshold from signal response measurements of a sample such that it is less than or equal to population decision threshold for 90% POD. Sampling error implies that this NDE reliability condition is violated. One of sampling types is called a representative sample. Representative samples reduce variance in POD estimates but also reduce magnitude of the error. Sampling sensitivity analysis for some sampling types is performed here using repetitive random sampling or Monte Carlo method. Six sampling types are considered for comparison. Some of the sampling types are similar to drawing a representative sample. LS POD model assumes random sampling. Therefore, random sampling is used as a basis for comparison with each sampling type. The sampling types used in the analysis are, A. Nominal and worst-case sampling, B. Worst-case sampling, C. Nominal case sampling, D. Random sampling, E. Random target, and sub-target sampling. F. Nominal target and sub-target sampling. Results of Monte Carlo simulation indicate that type F sampling can mitigate sampling risk and is also more practical to implement. Type A sampling may also mitigate the sampling risk, but it may be less practical to implement.

Ajay M Koshti↗

Characterization of electron microscopes with binary pseudo-random multilayer test samples

We discuss the results of SEM and TEM measurements with the BPRML test samples fabricated from a BPRML (WSi2/Si with fundamental layer thickness of 3 nm) with a Dual Beam FIB (focused ion beam)/SEM technique. In particular, we demonstrate that significant information about the metrological reliability of the TEM measurements can be extracted even when the fundamental frequency of the BPRML sample is smaller than the Nyquist frequency of the measurements. The measurements demonstrate a number of problems related to the interpretation of the SEM and TEM data. Note that similar BPRML test samples can be used to characterize x-ray microscopes. Corresponding work with x-ray microscopes is in progress.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

A new efficient grain growth model using a random Gaussian-sampled mode filter

This paper presents the use of a Gaussian neighborhood mode filter for predicting grain growth in a manner similar to the solutions obtained by a Monte Carlo Potts model. This flexible grain growth model can quickly utilize modern, computationally optimized data science strategies on graphics processing units to simulate grain growth up to 100 times faster than the state-of-the-art, publicly available Monte Carlo Potts model. We show that, given the correct neighborhood, the mode filter can replicate normal grain growth in two or three dimensions. In addition, the paper briefly demonstrates the ability to model limited anisotropic in grain boundary energy and mobility. Anisotropic grain boundary energy is modeled by defining a weighted mode filter operation. Anisotropic grain boundary mobility is modeled by scaling and orienting the Gaussian neighborhood in a particular direction.

Anisotropy↗

Visual Sample Plan (VSP) Statistical Software as Related to the CTBTO’s On-Site Inspection Procedure

In the event of a potential nuclear weapons test the Comprehensive Nuclear-Test-Ban Treaty Organization (CTBTO) is commissioned to conduct an on-site investigation (OSI) of the suspected test site in an effort to find confirmatory evidence of the nuclear test. The OSI activities include collecting air, surface soil, and underground samples to search for indications of a nuclear weapons test - these indicators include radionuclides and radioactive isotopes Ar and Xe. This report investigates the capability of the Visual Sample Plan (VSP) software to contribute to the sampling activities of the CTBTO during an OSI. VSP is a statistical sampling design software, constructed under data quality objectives, which has been adapted for environmental remediation and contamination detection problems for the EPA, US Army, DoD and DHS among others. This report provides discussion of a number of VSP sample designs, which may be pertinent to the work undertaken during an OSI. Examples and descriptions of such designs include hot spot sampling, combined random and judgment sampling, multiple increment sampling, radiological transect surveying, and a brief description of other potentially applicable sampling methods. Further, this work highlights a potential need for the use of statistically based sample designs in OSI activities. The use of such designs may enable canvassing a sample area without full sampling, provide a measure of confidence that radionuclides are not present, and allow investigators to refocus resources in other areas of concern.

98 NUCLEAR DISARMAMENT, SAFEGUARDS, AND PHYSICAL P↗

Eco-Friendly High-Performance Carbon Building Material Development from Coal

The development of eco-friendly high-performance building materials based on coal-derived materials is greatly beneficial for promoting environmental sustainability and the coal industry. This project focused on advancing technology to create innovative and scalable construction materials from two domestic sources: (1) coal-derived pyrolyzed char (PC), and (2) solvent-extracted coal deposit, extracts, and residue (CDER). It aimed to simultaneously develop and produce two types of coal-derived building products: (1) char-based concrete bricks (CCB) for wall construction in buildings, and (2) carbon-based structural units (CSU) intended as alternatives to traditional wood, concrete, or steel framing structures. The project team has successfully developed methods for manufacturing CCB samples. Extensive experiments were conducted to assess various manufacturing techniques for these samples. This included measuring their physical properties such as thermal conductivity (ranging from 0.26 to 0.38 W/mK) and bulk density (ranging from 0.85 to 1.15 g/cm 3 ), along with their mechanical properties (e.g., compressive strength ranging from 14 to 16 MPa at 28 days) for samples with a PC composition of 70%. Quality assurance and quality control tests, focusing on density and compressive strength, were also performed on samples selected randomly from pilot-scale production, and cured for a long term (over three months). The average long-term compressive strength of two randomly selected CCB samples was measured as 15.1 MPa, which was higher than the required target strength of 14 MPa. This indicates that the manufactured CCB samples from the preliminary pilot scale production run have good quality control and long-term strength performance.

01 COAL, LIGNITE, AND PEAT↗

Oscillating-flow regenerator test rig: Woven screen and metal felt results

We present correlating expressions, in terms of Reynolds or Peclet numbers, for friction factors, Nusselt numbers, enhanced axial conduction ratios, and overall heat flux ratios in four porous regenerator samples representative of stirling cycle regenerators: two woven screen samples and two random wire samples. Error estimates and comparison of data with others suggest our correlations are reliable, but we need to test more samples over a range of porosities before our results will become generally useful.

Gedeon, D.↗

Benchmarking Bayesian Optimization Frameworks and Acquisition Strategies for Materials Discovery and Autonomous Laboratories

Bayesian optimization (BO) can accelerate materials discovery by guiding expensive experiments toward the most promising processing conditions. We systematically compare five BO surrogate and framework combinations (Gaussian processes in Ax, Gaussian processes and Monte-Carlo neural networks in BayBE, random forests in Lolopy, and tree-structured Parzen (TPE) estimators in Hyperopt) on three benchmarks that mimic common materials design tasks (a discrete solid-electrolyte composition space, a hybrid discrete/continuous laminate-composite design problem solved with micromechanics modeling, and the continuous Ishigami analytic function which is a standard optimization benchmark). Each BO surrogate is paired with posterior mean, probability of improvement, and expected improvement acquisition functions and run for 100 trials from randomized initial samples with uniform random search providing a control. Across five random seeds per setting, BayBE’s Gaussian-process surrogate with expected improvement consistently reached ≥95 % of the known optimum in the fewest evaluations, while Lolopy’s random forest matched or exceeded GP performance on purely categorical or mixed spaces at a higher computational cost. Posterior mean alone often stagnated at local optima, underscoring the need for exploration, whereas probability and expected improvement balanced exploration and exploitation leading to better optimization in fewer trials. Execution times ranged from milliseconds for TPE to minutes for neural-network and random-forest surrogates. These results establish baseline expectations for BO in automated materials laboratories and highlight expected improvement with Gaussian processes as a reliable first choice, with random forests offering a strong alternative when categorical variables dominate. The benchmark suite and code are released to facilitate future surrogate, acquisition, and constraint-handling research in data-driven materials optimization.

Bayesian optimization↗