Search NASA⌕ Search

SEARCH · Search NASA

Results for “feature set”

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 325 records · Page 18
All sourcesNASA NTRSDOE OSTI
NASA NTRS1997

Development of a Simulated Data Set for the SeaWiFS Mission

A realistic simulated data set is essential for mission readiness preparations and can potentially assist in all phases of ground support for a future mission. Such a data set was created for the Sea-viewing Wide Field-of-view Sensor (SeaWiFS), a global ocean color mission due for launch in 1997. This data set incorporates a representation of virtually every known aspect of the flight mission. Thus, it provides a high fidelity data set for testing most phases of the ground system, Including data processing, data transfers, calibration and validation, quality control, and mission operations. The data set is constructed for a seven-day period, March 25-31, 1994. Specific features of the data set: it includes Global Area Coverage (GAC), recorded Local Area Coverage (LAC), and real-time High Resolution Picture Transmission (HRFIT) data for the seven-day period; it includes a realistic orbit which is propagated using a Brouwer-Lyddane model with drag; the data correspond to a command schedule based on the orbit for this seven-day period; it includes total (at-satellite) radiances for ocean, land, clouds, and ice; it utilizes a high-resolution land/sea mask; it includes actual SeaWiFS spectral responses; it includes the actual sensor saturation responses; it is formatted according to current onboard data structures; and it includes corresponding telemetry (instrument and spacecraft) data. The methods are described and some examples of the output are given.

Gregg, Watson W.↗
DOE OSTI2022

Predicting failure parameters of semiconductor devices subjected to stress conditions

A method for predicting failure parameters of semiconductor devices can include receiving a set of data that includes (i) characteristics of a sample semiconductor device, and (ii) parameters characterizing a stress condition. The method further includes extracting a plurality of feature values from the set of data and inputting the plurality of feature values into a trained model executing on the one or more processors, wherein the trained model is configured according to an artificial intelligence (AI) algorithm based on a previous plurality of feature values, and wherein the trained model is operable to output a failure prediction based on the plurality of feature values. Further, the method includes generating, via the trained model, a predicted failure parameter of the sample semiconductor device due to the stress condition.

Ahmed, Moinuddin↗
DOE OSTI2024

Exploiting Multi-Domain Features for Detection of Unclassified Electromagnetic Signals (Presentation)

Deep Learning based classification techniques have shown excellent performance in static environments, where the training and testing samples are drawn from the same distribution. However, real world scenarios often present samples that do not belong to the known set of classes chosen during training. This is quite common for electromagnetic signals, where it is impractical to assume that all possible waveforms are known a-priori, specially in scenarios like warfare. To address this problem, we propose a deep learning based adversarial model where the generator learns to generate waveform features that can deceive the discriminator model as true samples. We introduce domain knowledge of wireless signals by decomposing the signal into a lower dimensional unique feature set, which is used for classifying known versus unknown signals. We further introduce multiple domain representations of the signal to extract features and combine them together to accurately classify new waveforms as an unknown class. Our results show that combined features from multiple domains outperform any single domain representation, especially at low SNR regimes with fewer number of samples to classify.

99 - GENERAL AND MISCELLANEOUS↗
DOE OSTI2024

Exploiting Multi-Domain Features for Detection of Unclassified Electromagnetic Signals

Deep Learning based classification techniques have shown excellent performance in static environments, where the training and testing samples are drawn from the same distribution. However, real world scenarios often present samples that do not belong to the known set of classes chosen during training. This is quite common for electromagnetic signals, where it is impractical to assume that all possible waveforms are known a-priori, specially in scenarios like warfare. To address this problem, we propose a deep learning based adversarial model where the generator learns to generate waveform features that can deceive the discriminator model as true samples. We introduce domain knowledge of wireless signals by decomposing the signal into a lower dimensional unique feature set, which is used for classifying known versus unknown signals. We further introduce multiple domain representations of the signal to extract features and combine them together to accurately classify new waveforms as an unknown class. Our results show that combined features from multiple domains outperform any single domain representation, especially at low SNR regimes with fewer number of samples to classify.

99 - GENERAL AND MISCELLANEOUS↗
DOE OSTI2023

Exploratory analysis of machine learning techniques in the Nevada geothermal play fairway analysis

Play fairway analysis (PFA) is commonly used to generate geothermal potential maps and guide exploration studies, with a particular focus on locating and characterizing blind geothermal systems. This study evaluates the application of machine learning techniques to PFA in the Great Basin region of Nevada. Following the evaluation of various techniques, we identified two approaches to PFA that produced promising results, 1) supervised Bayesian probabilistic neural networks to generate geothermal potential maps with confidence intervals, and 2) unsupervised principal component analysis paired with k-means clustering to generate both cluster maps to help identify spatial patterns, as well as new combined feature inputs. We applied these techniques to perform a comparative analysis between two principal sets of geological and geophysical features related to permeability and heat and a set of positive (known geothermal resources) and negative training sites (known drill sites with unsuitable geothermal conditions). We found that these methods constrain previously unrecognized feature controls on geothermal favorability, many of which are spatially organized within the extent of cluster groups and the major structural-hydrologic domains of the study area. Furthermore, we utilized exploratory unsupervised modeling to highlight spatial relationships between input data and predictive output results of our supervised modeling. As a result, we demonstrate how our models compare to the previous Nevada PFA and how the rapid insights these machine learning techniques offer may support future assessments of both known and undiscovered blind geothermal systems in the Great Basin region of Nevada and beyond.

15 GEOTHERMAL ENERGY↗
DOE OSTI2023

Measurement report: Understanding the seasonal cycle of Southern Ocean aerosols

Abstract. The remoteness and extreme conditions of the Southern Ocean and Antarctic region have meant that observations in this region are rare, and typically restricted to summertime during research or resupply voyages. Observations of aerosols outside of the summer season are typically limited to long-term stations, such as Kennaook / Cape Grim (KCG; 40.7∘ S, 144.7∘ E), which is situated in the northern latitudes of the Southern Ocean, and Antarctic research stations, such as the Japanese operated Syowa (SYO; 69.0∘ S, 39.6∘ E). Measurements in the midlatitudes of the Southern Ocean are important, particularly in light of recent observations that highlighted the latitudinal gradient that exists across the region in summertime. Here we present 2 years (March 2016–March 2018) of observations from Macquarie Island (MQI; 54.5∘ S, 159.0∘ E) of aerosol (condensation nuclei larger than 10 nm, CN10) and cloud condensation nuclei (CCN at various supersaturations) concentrations. This important multi-year data set is characterised, and its features are compared with the long-term data sets from KCG and SYO together with those from recent, regionally relevant voyages. CN10 concentrations were the highest at KCG by a factor of ∼50 % across all non-winter seasons compared to the other two stations, which were similar (summer medians of 530, 426 and 468 cm−3 at KCG, MQI and SYO, respectively). In wintertime, seasonal minima at KCG and MQI were similar (142 and 152 cm−3, respectively), with SYO being distinctly lower (87 cm−3), likely the result of the reduction in sea spray aerosol generation due to the sea ice ocean cover around the site. CN10 seasonal maxima were observed at the stations at different times of year, with KCG and MQI exhibiting January maxima and SYO having a distinct February high. Comparison of CCN0.5 data between KCG and MQI showed similar overall trends with summertime maxima and wintertime minima; however, KCG exhibited slightly (∼10 %) higher concentrations in summer (medians of 158 and 145 cm−3, respectively), whereas KCG showed ∼40 % lower concentrations than MQI in winter (medians of 57 and 92 cm−3, respectively). Spatial and temporal trends in the data were analysed further by contrasting data to coincident observations that occurred aboard several voyages of the RSV Aurora Australis and the RV Investigator. Results from this study are important for validating and improving our models and highlight the heterogeneity of this pristine region and the need for further long-term observations that capture the seasonal cycles.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
NASA NTRS2000

Interactive Computer Graphics

Aerospace data analysis tools that significantly reduce the time and effort needed to analyze large-scale computational fluid dynamics simulations have emerged this year. The current approach for most postprocessing and visualization work is to explore the 3D flow simulations with one of a dozen or so interactive tools. While effective for analyzing small data sets, this approach becomes extremely time consuming when working with data sets larger than one gigabyte. An active area of research this year has been the development of data mining tools that automatically search through gigabyte data sets and extract the salient features with little or no human intervention. With these so-called feature extraction tools, engineers are spared the tedious task of manually exploring huge amounts of data to find the important flow phenomena. The software tools identify features such as vortex cores, shocks, separation and attachment lines, recirculation bubbles, and boundary layers. Some of these features can be extracted in a few seconds; others take minutes to hours on extremely large data sets. The analysis can be performed off-line in a batch process, either during or following the supercomputer simulations. These computations have to be performed only once, because the feature extraction programs search the entire data set and find every occurrence of the phenomena being sought. Because the important questions about the data are being answered automatically, interactivity is less critical than it is with traditional approaches.

Kenwright, David↗
NASA NTRS2004

Three Years of TRMM Precipitation Features: Radar, Radiometric, and Lightning Characteristics - Part 1

During its first three years, the Tropical Rainfall Measuring Mission (TRMM) satellite observed nearly six million precipitation features. The population of precipitation features is sorted by lightning flash rate, minimum brightness temperature, maximum radar reflectivity, areal extent, and volumetric rainfall. For each of these characteristics, essentially describing the convective intensity or the size of the features, the population is broken into categories consisting of the top 0.001%, top 0.01%, top 0.1%, top 1%, top 2.4%, and remaining 97.6%. The set of 'weakest / smallest' features comprises 97.6% of the population because that fraction does not have detected lightning, with a minimum detectable flash rate 0.7 fl/min. The greatest observed flash rate is 1351 fl/min; the lowest brightness temperatures are 42 K (85-GHz) and 69 K (37- GHz). The largest precipitation feature covers 335,000 sq km and the greatest rainfall from an individual precipitation feature exceeds 2 x 10(exp 12) kg of water. There is considerable overlap between the greatest storms according to different measures of convective intensity. The largest storms are mostly independent of the most intense storms. The set of storms producing the most rainfall is a convolution of the largest and the most intense storms. This analysis is a composite of the global tropics and subtropics. Significant variability is known to exist between locations, seasons, and meteorological regimes. Such variability will be examined in Part II. In Part I, only a crude land / Ocean separation is made. The known differences in bulk lightning flash rates over land and Ocean result from at least two differences in the precipitation feature population: the frequency of occurrence of intense storms, and the magnitude of those intense storms that do occur. Even when restricted to storms with the same brightness temperature, same size, or same radar reflectivity aloft, the storms over water are considerably less likely to produce lightning than are comparable storms over land.

Cecil, Daniel J.↗
DOE OSTI2020

Uncertainty Quantification of Machine Learning Predicted Creep Property of Alumina-Forming Austenitic Alloys

The development of machine learning (ML) approaches in materials science offers the opportunity to exploit existing engineering and developmental alloy datasets, such as Oak Ridge National Laboratory (ORNL)’s consistently measured creep-rupture dataset for alumina-forming austenitic (AFA) alloys, to accelerate their further development. As a first step toward achieving ML insights for improved alloy design, the potential sources of uncertainty and their impacts on ML output are examined. It is observed that the selection of algorithms and features as well as data sampling significantly affects the performance of ML models, either positively or negatively. Further, the performance of various ML models in predicting the creep properties of AFA alloys is compared, with further evaluation by assessment of a small set of new developmental AFA alloys that were not part of the training dataset. The present study demonstrates that uncertainty quantification (UQ) is essential in materials science for evaluating the performance of ML algorithms with specifically selected feature sets and obtaining a comprehensive understanding of their limitations and the resultant capability of effective prediction in complex materials systems.

36 MATERIALS SCIENCE↗
NASA NTRS1988

Geological remote sensing signatures of terrestrial impact craters

Geological remote sensing techniques can be used to investigate structural, depositional, and shock metamorphic effects associated with hypervelocity impact structures, some of which may be linked to global Earth system catastrophies. Although detailed laboratory and field investigations are necessary to establish conclusive evidence of an impact origin for suspected crater landforms, the synoptic perspective provided by various remote sensing systems can often serve as a pathfinder to key deposits which can then be targetted for intensive field study. In addition, remote sensing imagery can be used as a tool in the search for impact and other catastrophic explosion landforms on the basis of localized disruption and anomaly patterns. In order to reconstruct original dimensions of large, complex impact features in isolated, inaccessible regions, remote sensing imagery can be used to make preliminary estimates in the absence of field geophysical surveys. The experienced gained from two decades of planetary remote sensing of impact craters on the terrestrial planets, as well as the techniques developed for recognizing stages of degradation and initial crater morphology, can now be applied to the problem of discovering and studying eroded impact landforms on Earth. Preliminary results of remote sensing analyses of a set of terrestrial impact features in various states of degradation, geologic settings, and for a broad range of diameters and hence energies of formation are summarized. The intention is to develop a database of remote sensing signatures for catastrophic impact landforms which can then be used in EOS-era global surveys as the basis for locating the possibly hundreds of missing impact structures. In addition, refinement of initial dimensions of extremely recent structures such as Zhamanshin and Bosumtwi is an important objective in order to permit re-evaluation of global Earth system responses associated with these types of events.

Garvin, J. B.↗
DOE OSTI2020

A neural network for determination of latent dimensionality in Nonnegative Matrix Factorization

Non-negative Matrix Factorization (NMF) has proven to be a powerful unsupervised learning method for uncovering hidden features in complex and noisy datasets with applications in data mining, text recognition, dimension reduction, face recognition, anomaly detection, blind source separation, and many other fields. An important input for NMF is the latent dimensionality of the data, that is, the number of hidden features, K, present in the explored dataset. Unfortunately, and this quantity is rarely known a priori. The existing methods for determining latent dimensionality, such as Automatic Relevance Determination (ARD), are mostly heuristic and utilize different characteristics to estimate the number of hidden features. However, all of them require human presence to make a final determination of K. Here we utilize a supervised machine learning approach in combination with a recent method for model determination, called NMFk, to determine the number of hidden features automatically. NMFk performs a set of NMF simulations on an ensemble of matrices, obtained by bootstrapping the initial dataset, and estimates which K produces stable groups of latent features that reconstruct the initial dataset well. We then train a Multi-Layter Perceptron (MLP) classifier network to determine the correct number of latent features utilizing the statistics and characteristics of the NMF solution, obtained from NMFk. In order to train the MLP classifier, a training set of 58,660 matrices with predetermined latent features were factorized with NMFk. The MLP classifier in conjunction with NMFk maintains a greater than 95% success rate when applied to a held out test set. Additionally, when applied to two well-known benchmark datasets, the swimmer and MIT face data, NMFk/MLP correctly recovers the established number of hidden features. Finally, we compare the accuracy of our method to the ARD, AIC and Stability-based methods.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Decoding defect statistics from diffractograms via machine learning

Abstract Diffraction techniques can powerfully and nondestructively probe materials while maintaining high resolution in both space and time. Unfortunately, these characterizations have been limited and sometimes even erroneous due to the difficulty of decoding the desired material information from features of the diffractograms. Currently, these features are identified non-comprehensively via human intuition, so the resulting models can only predict a subset of the available structural information. In the present work we show (i) how to compute machine-identified features that fully summarize a diffractogram and (ii) how to employ machine learning to reliably connect these features to an expanded set of structural statistics. To exemplify this framework, we assessed virtual electron diffractograms generated from atomistic simulations of irradiated copper. When based on machine-identified features rather than human-identified features, our machine-learning model not only predicted one-point statistics (i.e. density) but also a two-point statistic (i.e. spatial distribution) of the defect population. Hence, this work demonstrates that machine-learning models that input machine-identified features significantly advance the state of the art for accurately and robustly decoding diffractograms.

36 MATERIALS SCIENCE↗
DOE OSTI2023

Temperature Dependence of Band Gap Renormalization in High-T Sensor Materials via First-Principles and Experimental Corroboration

Understanding the temperature dependence of functional properties of high-T gas sensing materials is vital for their applications in combustion environments. The electron-phonon coupling that derives the electronic structure change with temperatures is a key property of interest as it affects other sensing responses. Herein, we assess the temperature dependence of band gap renormalization in metal oxides and perovskites by employing Allen-Heine-Cardona theory with first-principles simulations and corroborate with experimental observation. The calculated temperature-dependent band gap changes of these materials studied are in good agreement with in-house experimental data, proving that the theory can adequately predict renormalization on the band gap in the system of interest. The predicted and measured band gap variations are characterized using an analytical model, which can provide useful insights on the simulated zero-temperature band gaps. Based on the available data, a set of 53 metal oxides and perovskites were identified as potential high-T gas sensors. A machine learning model has been developed to predict the band-gap change by capturing the overall trend of the empirical parameters with respect to a reduced feature obtained by transforming the set of available physical features.

Park, Jongwoo↗
NASA NTRS1991

Perceiving the coherent movements of spatially-separated features

When a partially-occluded object is represented in an image, it is defined by a set of spatially-separated features that may be registered at different spatial scales. To understand the image, human vision must organize these fragmented optical features into common and distinct object surfaces. Although the common fate of moving features often is considered a primary source of reliable information for image segmentation, little is known of the visual system's capacity to discriminate the coherent relative movements of spatially-separated features. In a series of experiments, observers viewed elements whose movements were correlated (direction and magnitude) or were uncorrelated. Our results indicate that observers can discriminate the two types of movement about as well as they can detect any movement at all. Moreover, the ability to perceive coherent movements is maintained under a variety of conditions including differences in the elements' spatial frequency content, spatial position and contrast, and temporal phase shifts between the spatially-correlated displacements. These results suggest that coherent relative motion may be a fundamental source of information exploited by vision, despite considerable variability in the spatial and temporal characteristics of the individual features.

Mowafy, Lyn↗
DOE OSTI2021

A novel machine learning based identification of potential adopter of rooftop solar photovoltaics

With the proliferation of rooftop solar photovoltaic installations, there is a need to proactively predict consumer potential for solar photovoltaic adoption, for improved electric utility planning and operation. Traditional analytical modeling approaches are limited to a few survey features and a larger part of the survey would remain untouched by the decision model. This article presents a novel, data-driven modeling approach that strategically prunes a large set of consumer profile features using a machine learning framework to train a model for predicting potential solar adoption. The approach utilizes the Gradient Boosting Decision Tree model through a Light Gradient Boosting framework that improves significantly over the poor prediction accuracy of the existing approaches. Model training using focal-loss based supervision is used to overcome the difficulty in identifying the potential adopters that is inherent in conventional data-driven models. In addition, to overcome possible data sparsity in a limited survey sample, a Generative Adversarial Network is presented to create synthetic user samples and its effectiveness on model performance is assessed. A Bayesian optimization approach is used to systematically arrive at the hyperparameters of the proposed model. Validation of the presented approach on a survey data collected by the National Rural Electric Cooperative Association in Virginia in 2018 demonstrates the excellent predictive capability of the machine learning based approach to modeling solar adoption reliably.

14 SOLAR ENERGY↗
NASA NTRS1981

Ultraviolet observations of the Io torus from the IUE observatory

The short wavelength spectrograph on the International Ultraviolet Explorer (IUE) has been used to obtain 11 A resolution spectra of the Io torus from 1175-1950 A. The four spectra, obtained in the springs of 1979 and 1980, show emissions (about 40R) of S II A1256A and S III A1199A. An unidentified feature is also present at 1729 A; a tentative identification as an intercombination line of S III is proposed. Weak features, probably due to O III A1664A and S IV A1406A, appear in some of the spectra. Abundances of the ions are determined from the brightnesses of the observed features. Upper limits are also set for the abundances of a number of ionic and neutral species. An observation of Io itself does not show any additional or enhanced spectral features, permitting upper limits to be set on the injection rate for a number of species.

Moos, H. W.↗
NASA NTRS1984

Analysis of the Gran Desierto, Pinacte Region, Sonora, Mexico, via shuttle imaging radar

The radar discriminability of geolian features and their geological setting as imaged by the SIR-A experiment is examined. The Gran Desierto and Pincate volcanio field of Sonora, Mexico was used to analyze the radar characteristics of the interplay of aeolian features and volcano terrain. The area in the Gran Desierto covers 4000 sq. km. and contains sand dunes of several forms. The Pincate volcanio field covers more than 2.000 sq. km. and consists primarily of basaltic lavas. Margins of the field, especially on the western and northern sides, include several maar and maar-like craters; thus obtaining information on their radar characteristics for comparison with impact craters.

Greeley, R.↗