Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “CLASSIFIED INFORMATION”

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 145 records · Page 8

Hierarchical image segmentation for learning object priors

The proposed segmentation approach naturally combines experience based and image based information. The experience based information is obtained by training a classifier for each object class. For a given test image, the result of each classifier is represented as a probability map. The final segmentation is obtained with a hierarchial image segmentation algorithm that considers both the probability maps and the image features such as color and edge strength. We also utilize image region hierarchy to obtain not only local but also semi-global features as input to the classifiers. Moreover, to get robust probability maps, we take into account the region context information by averaging the probability maps over different levels of the hierarchical segmentation algorithm. The obtained segmentation results are superior to the state-of-the-art supervised image segmentation algorithms.

47 OTHER INSTRUMENTATION↗

A review and evaluation of the Langley Research Center's Scientific and Technical Information Program. Results of phase 5. Design and evaluation of STI systems: A selected, annotated bibliography

A selected, annotated bibliography of literature citations related to the design and evaluation of STI systems is presented. The use of manual and machine-readable literature searches; the review of numerous books, periodicals reports, and papers; and the selection and annotation of literature citations were required. The bibliography was produced because the information was needed to develop the methodology for the review and evaluation project, and a survey of the literature did not reveal the existence of a single published source of information pertinent to the subject. Approximately 200 citations are classified in four subject areas. The areas include information - general; information systems - design and evaluation, including information products and services; information - use and need; and information - economics.

Pinelli, T. E.↗

Is infrared-collinear safe information all you need for jet classification?

Machine learning-based jet classifiers are able to achieve impressive tagging performance in a variety of applications in high-energy and nuclear physics. However, it remains unclear in many cases which aspects of jets give rise to this discriminating power, and whether jet observables that are tractable in perturbative QCD such as those obeying infrared-collinear (IRC) safety serve as sufficient inputs. In this article, we introduce a new classifier, Jet Flow Networks (JFNs), in an effort to address the question of whether IRC unsafe information provides additional discriminating power in jet classification. JFNs are permutation-invariant neural networks (deep sets) that take as input the kinematic information of reconstructed subjets. The subjet radius and a cut on the subjet’s transverse momenta serve as tunable hyperparameters enabling a controllable sensitivity to soft emissions and nonperturbative effects. We demonstrate the performance of JFNs for quark vs. gluon and Z vs. QCD jet tagging. For small subjet radii and transverse momentum cuts, the performance of JFNs is equivalent to the IRC-unsafe Particle Flow Networks (PFNs), demonstrating that infrared-collinear unsafe information is not necessary to achieve strong discrimination for both cases. As the subjet radius is increased, the performance of the JFNs remains essentially unchanged until physical thresholds that we identify are crossed. For relatively large subjet radii, we show that the JFNs may offer an increased model independence with a modest tradeoff in performance compared to classifiers that use the full particle information of the jet. These results shed new light on how machines learn patterns in high-energy physics data.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

A Machine Learns to Predict the Stability of Highly Diverse Multi-planetary Systems

Machine learning has proven to be an invaluable tool for characterizing the stability of planets in simplified planetary systems. In this work, we investigate the performance of a machine learning classifier on tightlypacked systems containing a rich diversity of planets, from Earths to Jupiters. Using information derived from short numerical simulations about a planet’s early orbital evolution and its relationship with the most massive planets in the system, we train a random forest classifier to predict instability with a > 88 percent accuracy. Our classifier relies on relative planet masses and the standard deviation of eccentricity for much of its predictive power. Most misclassified planets lie along a multi-dimensional boundary between stable and unstable planets, indicating that their early orbital evolution is ambiguous. The major reason for misclassification in this work is timescale: because our classifier uses information from only the first 137 years of simulation data, it is blind to late time interactions that cause or prevent instability. Machine learning methods like those utilized in this work provide powerful tools to complement numerical simulations across a wide range of planetary architectures.

79 ASTRONOMY AND ASTROPHYSICS↗

Seismic Event Characterization Using Full Moment Tensors on the Hypersphere

Moment tensor solutions provide insights into the deformation that has occurred in the source region of a seismic event and are therefore of great value in identifying different types of seismic sources, such as when monitoring for underground nuclear tests. Despite this utility, inversion of waveforms recorded by seismometers for their full seismic moment tensor is not yet routine, and development of robust methods to classify events based on this information is in its infancy. Here, we assemble an inventory of 1405 full moment tensor solutions that include explosive, earthquake, and collapse events, and investigate the use of anisotropic probability distribution functions on the 5D hypersphere to discriminate between these sources. Using a Bayesian classifier, we obtain optimal success rates of 98.4% across all events and demonstrate that modification of the prior probabilities provides a natural way to alter the balance between not missing desirable events (such as explosions) versus misclassifying large numbers of undesired events (such as earthquakes). The approach is specifically designed to progress from traditional, bipolar event screening metrics to more generalized event identification across multiple types of seismic sources. Despite current databases containing insufficient numbers of events to definitively demonstrate at present, we also find intriguing evidence of subgroupings within individual source populations on the hypersphere, for example, between chemical and nuclear explosions, raising the potential possibility of discriminating between these event types in the future.

Geosciences↗

Statistical processing of Pioneer front film data, part 1

A program was constructed to read the data on impacts and positional information on Pioneer and to classify these events according to a number of different criteria. The program is flexible enough to permit the introduction of further criteria and additional classifications, should this appear desirable. Not all cards correspond to particle impacts on the Pioneer sensors, many are inserted only to supply Pioneer position information.

Wolf, H.↗

Fast automatic 3D liver segmentation based on a three-level AdaBoost-guided active shape model

Purpose: A robust, automatic, and rapid method for liver delineation is urgently needed for the diagnosis and treatment of liver disorders. Until now, the high variability in liver shape, local image artifacts, and the presence of tumors have complicated the development of automatic 3D liver segmentation. In this study, an automatic three-level AdaBoost-guided active shape model (ASM) is proposed for the segmentation of the liver based on enhanced computed tomography images in a robust and fast manner, with an emphasis on the detection of tumors. Methods: The AdaBoost voxel classifier and AdaBoost profile classifier were used to automatically guide three-level active shape modeling. In the first level of model initialization, fast automatic liver segmentation by an AdaBoost voxel classifier method is proposed. A shape model is then initialized by registration with the resulting rough segmentation. In the second level of active shape model fitting, a prior model based on the two-class AdaBoost profile classifier is proposed to identify the optimal surface. In the third level, a deformable simplex mesh with profile probability and curvature constraint as the external force is used to refine the shape fitting result. In total, three registration methods—3D similarity registration, probability atlas B-spline, and their proposed deformable closest point registration—are used to establish shape correspondence. Results: The proposed method was evaluated using three public challenge datasets: 3Dircadb1, SLIVER07, and Visceral Anatomy3. The results showed that our approach performs with promising efficiency, with an average of 35 s, and accuracy, with an average Dice similarity coefficient (DSC) of 0.94 ± 0.02, 0.96 ± 0.01, and 0.94 ± 0.02 for the 3Dircadb1, SLIVER07, and Anatomy3 training datasets, respectively. The DSC of the SLIVER07 testing and Anatomy3 unseen testing datasets were 0.964 and 0.933, respectively. Conclusions: The proposed automatic approach achieves robust, accurate, and fast liver segmentation for 3D CTce datasets. The AdaBoost voxel classifier can detect liver area quickly without errors and provides sufficient liver shape information for model initialization. The AdaBoost profile classifier achieves sufficient accuracy and greatly decreases segmentation time. These results show that the proposed segmentation method achieves a level of accuracy comparable to that of state-of-the-art automatic methods based on ASM.

60 APPLIED LIFE SCIENCES↗

Searching for Low-mass Population III Stars Disguised as White Dwarfs

It is uncertain whether or not low-mass Population III stars ever existed. While limits on the number density of Population III stars with M {sub *} ≈ 0.8 M {sub ⊙} have been derived, using Sloan Digital Sky Survey (SDSS) data, little is known about the occurrence of Population III stars at lower masses. In the absence of reliable parallaxes, the spectra of metal-poor main-sequence (MPMS) stars with M {sub *} ≲ 0.8 M {sub ⊙} can easily be confused with those of cool white dwarfs. To resolve this ambiguity, in this paper we present a classifier that differentiates between MPMS stars and white dwarfs, based on photometry and/or spectroscopy without the use of parallax information. We build and train our classifier using state-of-the-art theoretical spectra, and evaluate it on existing SDSS-based classifications for objects with reliable Gaia DR2 parallaxes. We then apply our classifier to a large catalog of objects with SDSS photometry and spectroscopy to search for MPMS candidates. We discover several previously unknown, extremely metal-poor (EMP) candidate stars, and recover numerous confirmed EMP stars already noted in the literature. We conclude that archival SDSS spectroscopy has already been exhaustively searched for EMP stars. We predict that the lowest-mass stars of primordial composition will have redder optical-to-infrared colors than cool white dwarfs at constant effective temperature, due to surface gravity-dependent collision-induced absorption from molecular hydrogen. We suggest that the application of our classifier to data produced by next-generation spectroscopic surveys will set stronger constraints on the number density of low-mass Population III stars in the Milky Way.

47 OTHER INSTRUMENTATION↗

Hybrid Collaborative Learning for Classification and Clustering in Sensor Networks

Traditionally, nodes in a sensor network simply collect data and then pass it on to a centralized node that archives, distributes, and possibly analyzes the data. However, analysis at the individual nodes could enable faster detection of anomalies or other interesting events as well as faster responses, such as sending out alerts or increasing the data collection rate. There is an additional opportunity for increased performance if learners at individual nodes can communicate with their neighbors. In previous work, methods were developed by which classification algorithms deployed at sensor nodes can communicate information about event labels to each other, building on prior work with co-training, self-training, and active learning. The idea of collaborative learning was extended to function for clustering algorithms as well, similar to ideas from penta-training and consensus clustering. However, collaboration between these learner types had not been explored. A new protocol was developed by which classifiers and clusterers can share key information about their observations and conclusions as they learn. This is an active collaboration in which learners of either type can query their neighbors for information that they then use to re-train or re-learn the concept they are studying. The protocol also supports broadcasts from the classifiers and clusterers to the rest of the network to announce new discoveries. Classifiers observe an event and assign it a label (type). Clusterers instead group observations into clusters without assigning them a label, and they collaborate in terms of pairwise constraints between two events [same-cluster (mustlink) or different-cluster (cannot-link)]. Fundamentally, these two learner types speak different languages. To bridge this gap, the new communication protocol provides four types of exchanges: hybrid queries for information, hybrid "broadcasts" of learned information, each specified for classifiers-to-clusterers, and clusterers-to-classifiers. The new capability has the potential to greatly expand the in situ analysis abilities of sensor networks. Classifiers seeking to categorize incoming data into different types of events can operate in tandem with clusterers that are sensitive to the occurrence of new kinds of events not known to the classifiers. In contrast to current approaches that treat these operations as independent components, a hybrid collaborative learning system can enable them to learn from each other.

Wagstaff, Kiri L.↗

End-to-End Jet Classification of Boosted Top Quarks with CMS Open Data

We describe a novel application of the end-to-end deep learning technique to the task of discriminating top quark-initiated jets from those originating from the hadronization of a light quark or a gluon. The end-to-end deep learning technique combines deep learning algorithms and low-level detector representation of the high-energy collision event. In this study, we use lowlevel detector information from the simulated CMS Open Data samples to construct the top jet classifiers. To optimize classifier performance we progressively add low-level information from the CMS tracking detector, including pixel detector reconstructed hits and impact parameters, and demonstrate the value of additional tracking information even when no new spatial structures are added. Relying only on calorimeter energy deposits and reconstructed pixel detector hits, the end-to-end classifier achieves a ROC-AUC score of 0.975±0.002 for the task of classifying boosted top quark jets. After adding derived track quantities, the classifier ROC-AUC score increases to 0.9824±0.0013, serving as the first performance benchmark for these CMS Open Data samples.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Exploring Beyond Earth's Atmosphere with Human-Machine Teams

NASA's highly successful Kepler Mission has revolutionized our understanding of the Galaxy. We now know that planets, even Earth-size planets in the habitable zone, are common. With the end of the Kepler Mission we now look to the future with the Transiting Exoplanet Survey Satellite (TESS) which will discover thousands of exoplanets in orbit around the brightest stars in the sky. In a two-year survey, TESS will perform an all-sky search of more than 200,000 stars for temporary drops in brightness caused by planetary transits. With Kepler and TESS, humanity is finally at the verge of studying the masses, sizes, densities, orbits, and atmospheres of a large cohort of small planets, including a sample of rocky worlds in the habitable zones of their host stars which may prove to host life. The massive data sets generated by Kepler and TESS must be meticulously combed for the weakest planetary signals every month. While a daunting and error-prone task for humans, this is an exciting opportunity for the breakthroughs recently seen in machine learning. Specifically, traditional methods for identifying planet transits require extensive data processing pipelines followed by extensive human vetting. This manual process risks loss of information due to the data processing and to inconsistency and biases due to individual human vetters. The latest advancements in machine learning will allow an objective classifier to minimize the losses of information and greatly lessen the burden on the human vetters, in addition to providing assessment of quality and score to each planet candidate, freeing the humans to concentrate on border cases and other more interesting investigations.

Smith, Jeffrey C.↗

Classification of rain regimes by the three-dimensional properties of reflectivity fields

An automated scheme to characterize precipitation echoes within small windows in the radar field is presented and applied to previously subjectively classified tropical rain cloud systems near Darwin, Australia. The classification parameters are (a) E(sub e), effective efficiency, as determined by cloud-top and cloud-base water vapor saturation mixing ratios; (b) BBF, brightband fraction, as determined by the fraction of the radar echo area in which the maximal reflectivity occurs within +/- 1.5 km of the 0 C isotherm level; and (c) del(sub r) Z, radial reflectivity gradients (dB/km). These classification criteria were applied to tropical rain cloud systems near Darwin, Australia, and to winter convective rain cloud systems in Israel. Both sets of measurements were made with nearly identical networks of C-band radars and rain gauge networks. The results of the application of these objective classification criteria to several independently predetermined rain regimes in Darwin have shown that better organized rain systems have smaller del(sub r) Z and larger BBF. Similarly, smaller del(sub r)Z and larger BBF were also observed from maritime rain cloud systems, as compared to continental rain cloud systems with the same degree of organization. Continental rain cloud system, regardless of their degree of organization, have larger depths, as expressed by E(sub e). The rainfall analyses presented in this study are based exclusively on rain gauge measurements, while radar information was used only to classify the individual gauge measurements.

Rosenfeld, Daniel↗

An Information Theory Approach on Deciding Spectroscopic Follow-ups

Classification and characterization of variable phenomena and transient phenomena are critical for astrophysics and cosmology. These objects are commonly studied using photometric time series or spectroscopic data. Given that many ongoing and future surveys are conducted in a time domain, and given that adding spectra provides further insights but requires more observational resources, it would be valuable to know which objects we should prioritize to have a spectrum in addition to a time series. We propose a methodology in a probabilistic setting that determines a priori which objects are worth taking a spectrum of to obtain better insights, where we focus on the insight of the type of the object (classification). Objects for which we query their spectrum are reclassified using their full spectral information. We first train two classifiers, one that uses photometric data and another that uses photometric and spectroscopic data together. Then for each photometric object we estimate the probability of each possible spectrum outcome. We combine these models in various probabilistic frameworks (strategies), which are used to guide the selection of follow-up observations. The best strategy depends on the intended use, whether it is obtaining more confidence or accuracy. For a given number of candidate objects (127, equal to 5% of the data set) for taking spectra, we improve the class prediction accuracy by 37% as opposed to 20% of a non-naive (non-random) best-baseline strategy. Our approach provides a general framework for follow-up strategies and can be extended beyond classification to include other forms of follow-ups beyond spectroscopy.

79 ASTRONOMY AND ASTROPHYSICS↗

On Feature, Classifier and Detector Fusers for 235U Signatures Using Gamma Spectral Counts

Three types of information fusion strategies are studied to assess the performance of classifiers for detecting low-level 235 U radiation sources, using features obtained from gamma spectra of NaI detectors. These three strategies are based on using two spectral region features, fusing eight classifiers of diverse designs, and fusing multiple detectors located at different positions around the source. The inner, middle and outer groups of detectors, within a formation of two concentric circles and a spiral of 21 detectors, are identified based on their distance to the source, which is located at the center. This study provides two main qualitative insights into this classification task. First, the fusion of detectors leads to an overall improved classification performance, least in the inner group, most in the outer group, and in between for the middle group. Second, several classifiers and fusers achieve lower training error which does not translate to lower generalization error, indicating their over-fitting to training data.

Rao, Nageswara↗

Selecting a Classification Ensemble and Detecting Process Drift in an Evolving Data Stream

We characterize the commercial behavior of a group of companies in a common line of business using a small ensemble of classifiers on a stream of records containing commercial activity information. This approach is able to effectively find a subset of classifiers that can be used to predict company labels with reasonable accuracy. Performance of the ensemble, its error rate under stable conditions, can be characterized using an exponentially weighted moving average (EWMA) statistic. The behavior of the EWMA statistic can be used to monitor a record stream from the commercial network and determine when significant changes have occurred. Results indicate that larger classification ensembles may not necessarily be optimal, pointing to the need to search the combinatorial classifier space in a systematic way. Results also show that current and past performance of an ensemble can be used to detect when statistically significant changes in the activity of the network have occurred. The dataset used in this work contains tens of thousands of high level commercial activity records with continuous and categorical variables and hundreds of labels, making classification challenging.

Ensemble classifiers↗

Prospects for fundamental physics with LISA

In this paper, which is of programmatic rather than quantitative nature, we aim to further delineate and sharpen the future potential of the LISA mission in the area of fundamental physics. Given the very broad range of topics that might be relevant to LISA,we present here a sample of what we view as particularly promising fundamental physics directions. We organize these directions through a “science-first” approach that allows us to classify how LISA data can inform theoretical physics in a variety of areas. For each of these theoretical physics classes, we identify the sources that are currently expected to provide the principal contribution to our knowledge, and the areas that need further development. Overall, the classification presented here should not be thought of as cast in stone, but rather as a fluid framework that is amenable to change with the flow of new insights in theoretical physics.

79 ASTRONOMY AND ASTROPHYSICS↗

Relational Neural Evolution Approach to Bank Failure Prediction

Relational neural networks as a concept offers a unique opportunity for improving classification accuracy by exploiting relational structure in data. The premise is that a relational classification technique, which uses information implicit in relationships, should classify more accurately than techniques that only examine objects in isolation. In this paper, we study the use of relational neural networks for predicting bank failure. Alongside classical financial ratios normally used as predictor variables, we introduced new relational variables for the network. The relational neural network structure, specified as a combination of feed forward and recurrent neural networks, is determined by bank data through neuro-evolution. We discuss empirical results comparing performance of the relational approach to standard propositional methods used for bank failure prediction.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Augmenting Satellite Precipitation Estimation with Lightning Information

We have used lightning information to augment the Precipitation Estimation from Remotely Sensed Imagery using an Artificial Neural Network - Cloud Classification System (PERSIANN-CCS). Co-located lightning data are used to segregate cloud patches, segmented from GOES-12 infrared data, into either electrified (EL) or non-electrified (NEL) patches. A set of features is extracted separately for the EL and NEL cloud patches. The features for the EL cloud patches include new features based on the lightning information. The cloud patches are classified and clustered using self-organizing maps (SOM). Then brightness temperature and rain rate (T-R) relationships are derived for the different clusters. Rain rates are estimated for the cloud patches based on their representative T-R relationship. The Equitable Threat Score (ETS) for daily precipitation estimates is improved by almost 12% for the winter season. In the summer, no significant improvements in ETS are noted.

artificial neural network↗