SEARCH · Engineering Papers

Results for “Classification bias”

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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 91 records · Page 5

All sources NASA NTRS DOE OSTI

NASA NTRS2016

Psychophysiological Sensing and State Classification for Attention Management in Commercial Aviation

Attention-related human performance limiting states (AHPLS) can cause pilots to lose airplane state awareness (ASA), and their detection is important to improving commercial aviation safety. The Commercial Aviation Safety Team found that the majority of recent international commercial aviation accidents attributable to loss of control inflight involved flight crew loss of airplane state awareness, and that distraction of various forms was involved in all of them. Research on AHPLS, including channelized attention, diverted attention, startle / surprise, and confirmation bias, has been recommended in a Safety Enhancement (SE) entitled "Training for Attention Management." To accomplish the detection of such cognitive and psychophysiological states, a broad suite of sensors has been implemented to simultaneously measure their physiological markers during high fidelity flight simulation human subject studies. Pilot participants were asked to perform benchmark tasks and experimental flight scenarios designed to induce AHPLS. Pattern classification was employed to distinguish the AHPLS induced by the benchmark tasks. Unimodal classification using pre-processed electroencephalography (EEG) signals as input features to extreme gradient boosting, random forest and deep neural network multiclass classifiers was implemented. Multi-modal classification using galvanic skin response (GSR) in addition to the same EEG signals and using the same types of classifiers produced increased accuracy with respect to the unimodal case (90 percent vs. 86 percent), although only via the deep neural network classifier. These initial results are a first step toward the goal of demonstrating simultaneous real time classification of multiple states using multiple sensing modalities in high-fidelity flight simulators. This detection is intended to support and inform training methods under development to mitigate the loss of ASA and thus reduce accidents and incidents.

Harrivel, Angela R.↗

DOE OSTI2020

First Cosmology Results using Supernovae Ia from the Dark Energy Survey: Survey Overview, Performance, and Supernova Spectroscopy

We present details on the observing strategy, data-processing techniques, and spectroscopic targeting algorithms for the first three years of operation for the Dark Energy Survey Supernova Program (DES-SN). This five-year program using the Dark Energy Camera mounted on the 4 m Blanco telescope in Chile was designed to discover and follow supernovae (SNe) Ia over a wide redshift range (0.05 < z < 1.2) to measure the equation-of-state parameter of dark energy. We describe the SN program in full: strategy, observations, data reduction, spectroscopic follow-up observations, and classification. From three seasons of data, we have discovered 12,015 likely SNe, 308 of which have been spectroscopically confirmed, including 251 SNe Ia over a redshift range of 0.017 < z < 0.85. We determine the effective spectroscopic selection function for our sample and use it to investigate the redshift-dependent bias on the distance moduli of SNe Ia we have classified. The data presented here are used for the first cosmology analysis by DES-SN (“DES-SN3YR”), the results of which are given in Dark Energy Survey Collaboration et al. The 489 spectra that are used to define the DES-SN3YR sample are publicly available at https://des.ncsa.illinois.edu/releases/sn.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2020

Understanding and Explicitly Measuring Linguistic and Stylistic Properties of Deception via Generation and Translation

Massive digital disinformation is one of the main risks of modern society. Hundreds of models and linguistic analyses have been done to compare and contrast misleading and credible content online. However, most models do not remove the confounding factor of a topic or narrative when training, so the resulting models learn a clear topical separation for misleading versus credible content. We study the feasibility of using two strategies to disentangle the topic bias from the models to understand and explicitly measure linguistic and stylistic properties of content from misleading versus credible content. First, we develop conditional generative models to create news content that is characteristic of different credibility levels. We perform multi-dimensional evaluation of model performance on mimicking both the style and linguistic differences that distinguish news of different credibility using machine translation metrics and classification models. We show that even though generative models are able to imitate both the style and language of the original content, additional conditioning on both the news category and the topic leads to reduced performance. In a second approach, we perform deception style ``transfer" by translating deceptive content into the style of credible content and vice versa. Extending earlier studies, we demonstrate that, when conditioned on a topic, deceptive content is shorter, less readable, more biased, and more subjective than credible content, and transferring the style from deceptive to credible content is more challenging than the opposite direction.

Saldanha, Emily G.↗

DOE OSTI2023

Using Host Galaxy Photometric Redshifts to Improve Cosmological Constraints with Type Ia Supernovae in the LSST Era

We perform a rigorous cosmology analysis on simulated Type Ia supernovae (SNe Ia) and evaluate the improvement from including photometric host galaxy redshifts compared to using only the "z spec " subset with spectroscopic redshifts from the host or SN. We use the Deep Drilling Fields (~50 deg 2 ) from the Photometric LSST Astronomical Time-Series Classification Challenge (PLAsTiCC) in combination with a low-z sample based on Data Challenge2. The analysis includes light-curve fitting to standardize the SN brightness, a high-statistics simulation to obtain a bias-corrected Hubble diagram, a statistical+systematics covariance matrix including calibration and photo-z uncertainties, and cosmology fitting with a prior from the cosmic microwave background. Compared to using the z spec subset, including events with SN+host photo-z results in (i) more precise distances for z > 0.5, (ii) a Hubble diagram that extends 0.3 further in redshift, and (iii) a 50% increase in the Dark Energy Task Force figure of merit (FoM) based on the w 0 w a CDM model. Analyzing 25 simulated data samples, the average bias on w 0 and w a is consistent with zero. The host photo-z systematic of 0.01 reduces FoM by only 2% because (i) most z < 0.5 events are in the z spec subset, (ii) the combined SN+host photo-z has ×2 smaller bias, and (iii) the anticorrelation between fitted redshift and color self-corrects distance errors. To prepare for analyzing real data, the next SN Ia cosmology analysis with photo-zs should include non–SN Ia contamination and host galaxy misassociations.

79 ASTRONOMY AND ASTROPHYSICS↗

NASA NTRS1997

Analysis of Texture Using the Fractal Model

Properties such as the fractal dimension (FD) can be used for feature extraction and classification of regions within an image. The FD measures the degree of roughness of a surface, so this number is used to characterize a particular region, in order to differentiate it from another. There are two basic approaches discussed in the literature to measure FD: the blanket method, and the box counting method. Both attempt to measure FD by estimating the change in surface area with respect to the change in resolution. We tested both methods but box counting resulted computationally faster and gave better results. Differential Box Counting (DBC) was used to segment a collage containing three textures. The FD is independent of directionality and brightness so five features were used derived from the original image to account for directionality and gray level biases. FD can not be measured on a point, so we use a window that slides across the image giving values of FD to the pixel on the center of the window. Windowing blurs the boundaries of adjacent classes, so an edge-preserving, feature-smoothing algorithm is used to improve classification within segments and to make the boundaries sharper. Segmentation using DBC was 90.8910 accurate.

Navas, William↗

DOE OSTI2022

Weakly Supervised Event Classification Using Imperfect Real-world PMU Data with Scarce Labels

This paper studies event classification using imperfect real-world phasor measurement unit (PMU) data with scarce event types (labels). By investigating the real-world PMU data, it is observed that most real-world PMU data's event type is unknown, which makes it challenging to directly use such dataset to build event classifiers as existing classification techniques require high-quality training data with known event type (i.e., label). To address this challenge, a weakly supervised learning based event classification approach is developed, which can use noisy and low-quality PMU data for the training. First, data quality issues are fixed using data preprocessing techniques and then event features are constructed from the PMU data. Using these features, a series of labeling functions are learnt to generate initial estimates of the labels of large amounts of unlabeled PMU data. As the labeling functions are learnt using the same data with scarce labels, the label estimates from the labeling functions can be correlated, noisy, and bias. To enhance these initial estimates, a generative model is developed to characterize the dependencies among the estimated labels, based on which better labels are obtained for training event classifiers. Numerical experiments using the real-world dataset from the Western Interconnection of the U.S. power transmission grid show that the proposed weakly supervised event classifier trained using the dataset with only 5% labeled data can achieve 78.4% classification accuracy.

Liu, Yunchuan↗

NASA NTRS2004

Toward a Global Map of Raindrop Size Distributions: Rain-Type Classification and Its Implications for Validating Global Rainfall Products - Part 1

Variability in the global distribution of precipitation is recognized as a key element in assessing the impact of climate change for life on earth. The response of precipitation to climate forcings is, however, poorly understood because of discrepancies in the magnitude and sign of climatic trends in satellite-based rainfall estimates. Quantifying and ultimately removing these biases is critical for studying the response of the hydrologic cycle to climate change. In addition, estimates of random errors owing to variability in algorithm assumptions on local spatial and temporal scales are critical for establishing how strongly their products should be weighted in data assimilation or model validation applications and for assigning a level of confidence to climate trends diagnosed from the data. This paper explores the potential for refining assumed drop size distributions (DSDs) in global radar rainfall algorithms by establishing a link between satellite observables and information gleaned from regional validation experiments where polarimetric radar, Doppler radar, and disdrometer measurements can be used to infer raindrop size distributions. By virtue of the limited information available in the satellite retrieval framework, the current method deviates from approaches adopted in the ground-based radar community that attempt to relate microphysical processes and resultant DSDs to local meteorological conditions. Instead, the technique exploits the fact that different microphysical pathways for rainfall production are likely to lead to differences in both the DSD of the resulting raindrops and the three-dimensional structure of associated radar reflectivity profiles. Objective rain-type classification based on the complete three-dimensional structure of observed reflectivity profiles is found to partially mitigate random and systematic errors in DSDs implied by differential reflectivity measurements. In particular, it is shown that vertical and horizontal reflectivity structure obtained from spaceborne radar can be used to reproduce significant differences in Z(sub dr) between the easterly and westerly climate regimes observed in the Tropical Rainfall Measuring Mission Large-scale Biosphere-Atmosphere (TRMM-LBA) field experiment as well as the even larger differences between Amazonian rainfall and that observed in eastern Colorado. As such, the technique offers a potential methodology for placing locally observed DSD information into a global framework.

L'Ecuyer, Tristan S.↗

DOE OSTI2020

Robust impurity detection and tracking for tokamaks

A robust impurity detection and tracking code, able to generate large sets of dust tracks from tokamak camera footage, is presented. This machine learning–based code is tested with cameras from the Joint European Torus, Doublet-III-D, and Magnum-PSI and is able to generate dust tracks with a 65 – 100 % classification accuracy. Moreover, the number dust particles detected from a single camera shot can be up to the order of 1000. Several areas of improvement for the code are highlighted, such as generating more significant training data sets and accounting for selection biases. Finally, although the code is tested with dust in single two-dimensional camera views, it could easily be applied to multiple-camera stereoscopic reconstruction or nondust impurities.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

NASA NTRS2002

How Sample Completeness Affects Gamma-Ray Burst Classification

Unsupervised pattern recognition algorithms support the existence of three gamma-ray burst classes; class I (long, large fluence bursts of intermediate spectral hardness), Class II (short, small fluence, hard bursts), and class III (soft bursts of intermediate durations and fluences). The algorithms surprisingly assign larger membership to class III than to either of the other two classes. A known systematic bias has been previously used to explain the existence of class III in terms of class I; this bias allows the fluences and durations of some bursts to be underestimated. We show that this bias primarily affects only the longest bursts and cannot explain the bulk of the class III properties. We resolve the question of class III existence by demonstrating how samples obtained using standard trigger mechanisms fail to preserve the duration characteristics of small peak flux bursts: (Sample incompleteness is thus primarily responsible for the existence of class III.) In order to avoid this incompleteness, we show how a new dual timescale peak flux can be defined in terms of peak flux and fluence. The dual timescale peak flux reserves the duration distribution of faint bursts and correlates either with spectral hardness (and presumably redshift) than either peak flux or fluence. The techniques presented here are generic and have applicability to the studies of other transient events. The results also indicate that pattern recognition algorithms are sensitive to sample completeness; this can influence the study of large astronomical databases such as those found in a Virtual Observatory.

Hakkila, Jon↗

NASA NTRS1993

Signal detection theory and methods for evaluating human performance in decision tasks

Signal Detection Theory (SDT) can be used to assess decision making performance in tasks that are not commonly thought of as perceptual. SDT takes into account both the sensitivity and biases in responding when explaining the detection of external events. In the standard SDT tasks, stimuli are selected in order to reveal the sensory capabilities of the observer. SDT can also be used to describe performance when decisions must be made as to the classification of easily and reliably sensed stimuli. Numbers are stimuli that are minimally affected by sensory processing and can belong to meaningful categories that overlap. Multiple studies have shown that the task of categorizing numbers from overlapping normal distributions produces performance predictable by SDT. These findings are particularly interesting in view of the similarity between the task of the categorizing numbers and that of determining the status of a mechanical system based on numerical values that represent sensor readings. Examples of the use of SDT to evaluate performance in decision tasks are reviewed. The methods and assumptions of SDT are shown to be effective in the measurement, evaluation, and prediction of human performance in such tasks.

Obrien, Kevin↗

NASA NTRS2013

Multitemporal Snow Cover Mapping in Mountainous Terrain for Landsat Climate Data Record Development

A multitemporal method to map snow cover in mountainous terrain is proposed to guide Landsat climate data record (CDR) development. The Landsat image archive including MSS, TM, and ETM+ imagery was used to construct a prototype Landsat snow cover CDR for the interior northwestern United States. Landsat snow cover CDRs are designed to capture snow-covered area (SCA) variability at discrete bi-monthly intervals that correspond to ground-based snow telemetry (SNOTEL) snow-water-equivalent (SWE) measurements. The June 1 bi-monthly interval was selected for initial CDR development, and was based on peak snowmelt timing for this mountainous region. Fifty-four Landsat images from 1975 to 2011 were preprocessed that included image registration, top-of-the-atmosphere (TOA) reflectance conversion, cloud and shadow masking, and topographic normalization. Snow covered pixels were retrieved using the normalized difference snow index (NDSI) and unsupervised classification, and pixels having greater (less) than 50% snow cover were classified presence (absence). A normalized SCA equation was derived to independently estimate SCA given missing image coverage and cloud-shadow contamination. Relative frequency maps of missing pixels were assembled to assess whether systematic biases were embedded within this Landsat CDR. Our results suggest that it is possible to confidently estimate historical bi-monthly SCA from partially cloudy Landsat images. This multitemporal method is intended to guide Landsat CDR development for freshwaterscarce regions of the western US to monitor climate-driven changes in mountain snowpack extent.

Landsat↗

NASA NTRS2003

How Sample Completeness Affects Gamma-Ray Burst Classification

Unsupervised pattern-recognition algorithms support the existence of three gamma-ray burst classes: class 1 (long, large-fluence bursts of intermediate spectral hardness), class 2 (short, small-fluence, hard bursts), and class 3 (soft bursts of intermediate durations and fluences). The algorithms surprisingly assign larger membership to class 3 than to either of the other two classes. A known systematic bias has been previously used to explain the existence of class 3 in terms of class 1 ; this bias allows the fluences and durations of some bursts to be underestimated, as recently shown by Hakkila et al. We show that this bias primarily affects only the longest bursts and cannot explain the bulk of the class 3 properties. We resolve the question of class 3's existence by demonstrating how samples obtained using standard trigger mechanisms fail to preserve the duration characteristics of small-peak flux bursts. Sample incompleteness is thus primarily responsible for the existence of class 3. In order to avoid this incompleteness, we show how a new, dual-timescale peak flux can be defined in terms of peak flux and fluence. The dual-timescale peak flux preserves the duration distribution of faint bursts and correlates better with spectral hardness (and presumably redshift) than either peak flux or fluence. The techniques presented here are generic and have applicability to the studies of other transient events. The results also indicate that pattern recognition algorithms are sensitive to sample completeness; this can influence the study of large astronomical databases, such as those found in a virtual observatory.

Hakkila, Jon↗

NASA NTRS1981

Evaluation of large area crop estimation techniques using LANDSAT and ground-derived data

The results of the Domestic Crops and Land Cover Classification and Clustering study on large area crop estimation using LANDSAT and ground truth data are reported. The current crop area estimation approach of the Economics and Statistics Service of the U.S. Department of Agriculture was evaluated in terms of the factors that are likely to influence the bias and variance of the estimator. Also, alternative procedures involving replacements for the clustering algorithm, the classifier, or the regression model used in the original U.S. Department of Agriculture procedures were investigated.

Amis, M. L.↗

NASA NTRSDate not supplied

Using Machine Learning to Infer Material Properties of Debris Fragments from X-ray Images in the DebriSat Project

The DebriSat project is a collaboration effort with the NASA Orbital Debris Program Office, the U.S. Space Force Space Systems Command Center, The Aerospace Corporation, and the University of Florida. To date, over 200,000 fragments from this ground-based, hypervelocity impact experiment have been collected, and processing is underway to determine their physical characteristics, such as material, shape, color, characteristic length, and average cross-sectional area. The x-ray process is primarily used to identify the location of the fragments and estimated size for extraction, so that these physical characteristics can be assessed. This paper proposes a machine learning-based approach to characterize materials from x-ray images of debris fragments embedded in soft-catch foam used in the DebriSat project. The novel methodology discussed in this paper will highlight the use of x-ray imagery data to characterize these fragments without extraction or a human-in-the-loop. Both supervised and unsupervised machine learning techniques are utilized with this approach to infer the physical parameters of the fragments embedded in the soft-catch foam panels used in the impact experiment based on x-ray images of the foam panels. Additionally, 3D reconstructions of the extracted fragments are created with images taken from two different angles using the structure from motion (SfM) method. The characteristic lengths and shape from the 3D reconstruction, alongside the physical characteristics of the debris, are used in the inference of the material type. To develop and test the approach, a dataset of x-ray images of debris fragments of varying sizes and materials is collected. Supervised learning methods such as convolutional neural networks (CNNs), support vector machines (SVM), decision trees, and random forest classifiers are used due to the high-dimensional feature spaces of the debris and nonlinear decision boundaries for material categorization. Given the limited pre-labeled data of embedded debris materials smaller than 10 mm, unsupervised machine learning techniques such as clustering algorithms and autoencoders are used, in addition to supervised learning methods. The clustering algorithms group similar fragments together based on their physical properties, and autoencoders reduce the dimensionality of the x ray images and extract relevant features. The performance of the proposed approach's is analyzed using a range of statistical methods, including confusion matrices, receiver operating characteristic curves, and precision-recall curves. The results are compared with those obtained using a baseline approach that relies on manual identification and classification of debris fragments. To evaluate the effectiveness of different machine learning methods, statistical tests such as t-tests, ANOVA, and cross-validation are performed, comparing the performance of CNNs, SVMs, clustering algorithms, and autoencoders. Additional analysis needs to be conducted to identify any sources of bias or variability that may affect the results, such as variations in imaging conditions or fragmentation patterns. Other topics explored are limitations, refinements, and the potential use of semi-supervised learning techniques, such as self-training to label unlabeled datasets and co-training using x-ray images taken from two different angles as two different models.

Saik Anam Siam↗

NASA NTRS2023

Using Machine Learning to Infer Material Properties of Debris Fragments from X-ray Images in the DebriSat Project

Saik Anam Siam↗

DOE OSTI2022

WISE-PS1-STRM: neural network source classification and photometric redshifts for WISE×PS1

ABSTRACT We cross-match between the WISE All-Sky and PS1 3π DR2 source catalogues. The resulting catalogue has 354 590 570 objects, significantly fewer than the parent PS1 catalogue, but its combination of optical and infrared colours facilitate both better source classification and photometric redshift estimation. We perform a neural network-based classification of the objects into galaxies, quasars, and stars, then run neural network-based photometric redshift estimation for the galaxies. The star sample purity and quasar sample completeness measures improve substantially, and the resulting photo-z’s are significantly more accurate in terms of statistical scatter and bias than those calculated from PS1 properties alone. The catalogue will be a basis for future large-scale structure studies, and will be made available as a high-level science product via the Mikulski Archive for Space Telescopes.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2021

Chemical Heterointerface Engineering on Hybrid Electrode Materials for Electrochemical Energy Storage

Abstract The chemical heterointerfaces in hybrid electrode materials play an important role in overcoming the intrinsic drawbacks of individual materials and thus expedite the in‐depth development of electrochemical energy storage. Benefiting from the three enhancement effects of accelerating charge transport, increasing the number of storage sites, and reinforcing structural stability, the chemical heterointerfaces have attracted extensive interest and the electrochemical performances of hybrid electrode materials have been significantly optimized. In this review, recent advances regarding chemical heterointerface engineering in hybrid electrode materials are systematically summarized. Especially, the intrinsic behaviors of chemical heterointerfaces on hybrid electrode materials are refined based on built‐in electric field, van der Waals interaction, lattice mismatch and connection, electron cloud bias and chemical bond, and their combination. The strategies for introducing chemical heterointerfaces are classified into in situ local transformation, in situ growth, cosynthesis, and other strategy. The recent progress about the chemical heterointerfaces engineering specially focusing on metal‐ion batteries, supercapacitors, and Li–S batteries are introduced in detail. Furthermore, the classification and characterization of chemical heterointerfaces are briefly described. Finally, the emerging challenges and perspectives about future directions of chemical heterointerface engineering are proposed.

Li, Wenbin↗

DOE OSTI2020

A map of roadmaps for zero and low energy and carbon buildings worldwide

Formulation of targets and establishing which factors in different contexts will achieve these targets are critical to successful decarbonization of the building sector. To contribute to this, we have performed an evidence map of roadmaps for zero and low energy and carbon buildings (ZLECB) worldwide, including a list and classification of documents in an on-line geographical map, a description of gaps, and a narrative review of the knowledge gluts. We have retrieved 1219 scientific documents from Scopus, extracted metadata from 274 documents, and identified 117 roadmaps, policies or plans from 27 countries worldwide. We find that there is a coverage bias towards more developed regions. The identified scientific studies are mostly recommendations to policy makers, different types of case studies, and demonstration projects. The geographical inequalities found in the coverage of the scientific literature are even more extreme in the coverage of the roadmaps. These underexplored world regions represent an area for further investigation and increased research/policy attention. Our review of the more substantial amount of literature and roadmaps for developed regions shows differences in target metrics and enforcement mechanisms but that all regions dedicate some efforts at national and local levels. Roadmaps generally focus more on new and public buildings than existing buildings, despite the fact that the latter are naturally larger in number and total floor area, and perform less energy efficiently. A combination of efficiency, technical upgrades, and renewable generation is generally proposed in the roadmaps, with behavioral measures only reflected in the use of information and communication technologies, and minimal focus being placed on lifecycle perspectives. We conclude that insufficient progress is being made in the implementation of ZLECB. More work is needed to couple the existing climate goals, with realistic, enforceable policies to make the carbon savings a reality for different contexts and stakeholders worldwide.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗