Search NASA⌕ Search

SEARCH · Search NASA

Results for “data reduction techniques”

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 307 records · Page 17
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Addressing GPU memory limitations for Graph Neural Networks in High-Energy Physics applications

Introduction Reconstructing low-level particle tracks in neutrino physics can address some of the most fundamental questions about the universe. However, processing petabytes of raw data using deep learning techniques poses a challenging problem in the field of High Energy Physics (HEP). In the Exa.TrkX Project, an illustrative HEP application, preprocessed simulation data is fed into a state-of-art Graph Neural Network (GNN) model, accelerated by GPUs. However, limited GPU memory often leads to Out-of-Memory (OOM) exceptions during training, due to the large size of models and datasets. This problem is exacerbated when deploying models on High-Performance Computing (HPC) systems designed for large-scale applications. Methods We observe a high workload imbalance issue during GNN model training caused by the irregular sizes of input graph samples in HEP datasets, contributing to OOM exceptions. We aim to scale GNNs on HPC systems, by prioritizing workload balance in graph inputs while maintaining model accuracy. Our paper introduces diverse balancing strategies aimed at decreasing the maximum GPU memory footprint and avoiding the OOM exception, across various datasets. Results Our experiments showcase memory reduction of up to 32.14% compared to the baseline. We also demonstrate the proposed strategies can avoid OOM in application. Additionally, we create a distributed multi-GPU implementation using these samplers to demonstrate the scalability of these techniques on the HEP dataset. Discussion By assessing the performance of these strategies as data loading samplers across multiple datasets, we can gauge their effectiveness in both single-GPU and distributed environments. Our experiments, conducted on datasets of varying sizes and across multiple GPUs, broaden the applicability of our work to various GNN applications that handle input datasets with irregular graph sizes.

Lee, Claire Songhyun↗
DOE OSTI2021

Preliminary results of neutron and X-ray diffraction data collection on a lytic polysaccharide monooxygenase under reduced and acidic conditions

Lytic polysaccharide monooxygenases (LPMOs) are copper-center enzymes that are involved in the oxidative cleavage of the glycosidic bond in crystalline cellulose and other polysaccharides. The LPMO reaction is initiated by the addition of a reductant and oxygen to ultimately form an unknown activated copper–oxygen species that is responsible for polysaccharide-substrate H-atom abstraction. Given the sensitivity of metalloproteins to radiation damage, neutron protein crystallo­graphy provides a nondestructive technique for structural characterization while also informing on the positions of H atoms. Neutron cryo-crystallography permits the trapping of catalytic intermediates, thereby providing insight into the protonation states and chemical nature of otherwise short-lived species in the reaction mechanism. Here, to characterize the reaction-mechanism intermediates of LPMO9D from Neurospora crassa, a cryo-neutron diffraction data set was collected from an ascorbate-reduced crystal. A second neutron diffraction data set was collected at room temperature from an LPMO9D crystal exposed to low-pH conditions to probe the protonation states of ionizable groups involved in catalysis under acidic conditions.

Neurospora crassa↗
DOE OSTI2020

A comparison of cost and quality of three methods for estimating density for wild pig (Sus scrofa)

A critical element in efective wildlife management is monitoring the status of wildlife populations; however, resources to monitor wildlife populations are typically limited. We compared cost efectiveness of three common population estimation methods (i.e. non-invasive DNA sampling, camera sampling, and sampling from trapping) by applying them to wild pigs (Sus scrofa) across three habitats in South Carolina, U.S.A where they are invasive. We used mark-recapture analyses for fecal DNA sampling data, spatially-explicit capture-recapture analyses for camera sampling data, and a removal analysis for removal sampling from trap data. Density estimates were similar across methods. Camera sampling was the least expensive, but had large variances. Fecal DNA sampling was the most expensive, although this technique generally performed well. We examined how reductions in efort by method related to increases in relative bias or imprecision. For removal sampling, the largest cost savings while maintaining unbiased density estimates was from reducing the number of traps. For fecal DNA sampling, a reduction in efort only minimally reduced costs due to the need for increased lab replicates while maintaining high quality estimates. For camera sampling, efort could only be marginally reduced before inducing bias. We provide a decision tree for researchers to help make monitoring decisions.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Unlocking the potential: machine learning applications in electrocatalyst design for electrochemical hydrogen energy transformation

Machine learning (ML) is rapidly emerging as a pivotal tool in the hydrogen energy industry for the creation and optimization of electrocatalysts, which enhance key electrochemical reactions like the hydrogen evolution reaction (HER), the oxygen evolution reaction (OER), the hydrogen oxidation reaction (HOR), and the oxygen reduction reaction (ORR). This comprehensive review demonstrates how cutting-edge ML techniques are being leveraged in electrocatalyst design to overcome the time-consuming limitations of traditional approaches. ML methods, using experimental data from high-throughput experiments and computational data from simulations such as density functional theory (DFT), readily identify complex correlations between electrocatalyst performance and key material descriptors. Leveraging its unparalleled speed and accuracy, ML has facilitated the discovery of novel candidates and the improvement of known products through its pattern recognition capabilities. This review aims to provide a tailored breakdown of ML applications in a format that is readily accessible to materials scientists. Hence, we comprehensively organize ML-driven research by commonly studied material types for different electrochemical reactions to illustrate how ML adeptly navigates the complex landscape of descriptors for these scenarios. We further highlight ML's critical role in the future discovery and development of electrocatalysts for hydrogen energy transformation. Potential challenges and gaps to fill within this focused domain are also discussed. As a practical guide, we hope this work will bridge the gap between communities and encourage novel paradigms in electrocatalysis research, aiming for more effective and sustainable energy solutions.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2011

CLASSIFICATION OF STELLAR SPECTRA WITH LOCAL LINEAR EMBEDDING

We investigate the use of dimensionality reduction techniques for the classification of stellar spectra selected from the Sloan Digital Sky Survey. Using local linear embedding (LLE), a technique that preserves the local (and possibly nonlinear) structure within high-dimensional data sets, we show that the majority of stellar spectra can be represented as a one-dimensional sequence within a three-dimensional space. The position along this sequence is highly correlated with spectral temperature. Deviations from this 'stellar locus' are indicative of spectra with strong emission lines (including misclassified galaxies) or broad absorption lines (e.g., carbon stars). Based on this analysis, we propose a hierarchical classification scheme using LLE that progressively identifies and classifies stellar spectra in a manner that requires no feature extraction and that can reproduce the classic MK classifications to an accuracy of one type.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2021

Accelerating Ultrafast Spectroscopy with Compressive Sensing

We report ultrafast spectroscopy is an important tool for studying photoinduced dynamical processes in atoms, molecules, and nanostructures. Typically, the time to perform these experiments ranges from several minutes to hours depending on the choice of spectroscopic method. It is desirable to reduce this time overhead not only to shorten time and laboratory resources, but also to make it possible to examine fragile specimens that quickly degrade during long experiments. In this article, we motivate using compressive sensing to significantly shorten data acquisition time by reducing the total number of measurements in ultrafast spectroscopy. We apply this technique to experimental data from ultrafast transient absorption spectroscopy and ultrafast terahertz spectroscopy and show that good estimates can be obtained with as low as 15% of the total measurements, implying a sixfold reduction in the data acquisition time.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
DOE OSTI2012

An Improved Technique for Increasing the Accuracy of Photometrically Determined Redshifts for "Blended" Galaxies

The redshift of a galaxy can be determined by one of two methods; photometric or spectroscopic. Photometric is a term for any redshift determination made using the magnitudes of light in different filters. Spectroscopic redshifts are determined by measuring the absorption spectra of the object then determining the difference in wavelength between the 'standard' absorption lines and the measured ones, making it the most accurate of the two methods. The data for this research was collected from SDSS DR8 and then separated into blended and non-blended galaxy sets; the definition of 'blended' is discussed in the Introduction section. The current SDSS photometric redshift determination method does not discriminate between blended and non-blended data when it determines the photometric redshift of a given galaxy. The focus of this research was to utilize machine learning techniques to determine if a considerably more accurate photometric redshift determination method could be found, for the case of the blended and non-blended data being treated separately. The results show a reduction of 0.00496 in the RMS error of photometric redshift determinations for blended galaxies and a more significant reduction of 0.00827 for non-blended galaxies, illustrated in Table 2.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2015

High-resolution neutron diffraction study of microstructural changes in nanocrystalline ball-milled niobium carbide NbC{sub 0.93}

High resolution neutron diffraction was applied for elucidating of the microstructural evolution of nanocrystalline niobium carbide NbC{sub 0.93} powders subjected to high-energy ball milling. The diffraction patterns were collected with the high resolution Fourier diffractometer HRFD by using the reverse time-of-flight (RTOF) mode of data acquisition. The traditional single diffraction line analysis, the Rietveld method and more advanced Whole Powder Pattern Modeling technique were applied for the data analysis. The comparison of these techniques was performed. It is established that short-time milling produces a non-uniform powder, in which two distinct fractions with differing microstructure can be identified. Part of the material is in fact milled efficiently, with a reduction in grain size, an increase in the quantity of defects, and a corresponding tendency to decarburize reaching a composition NbC{sub 0.80} after 15 h of milling. The rest of the powder is less efficiently processed and preserves its composition and lower defect content. Larger milling times should have homogenized the system by increasing the efficiently milled fraction, but the material is unable to reach a uniform and homogeneous state. It is definitely shown that RTOF neutron diffraction patterns can provide the very accurate data for microstructure analysis of nanocrystalline powders. - Highlights: • The NbC{sub 0.93} powder was processed by high-energy ball milling. • The microstrain and dislocation density increase with milling time increase. • The corresponding decrease in crystallite size with milling time was observed. • The material exhibits the presence of two fractions after ball milling. • The RTOF neutron diffraction data are suitable for accurate microstructure analysis.

36 MATERIALS SCIENCE↗
DOE OSTI2026

Deep-learning methods for contrast enhancement and artifact reduction in cryo-electron tomography: a systematic analysis of the state of the art and proposed improvements

Cryo-electron tomography (cryo-ET) has emerged as the preferred technique for visualizing the organization of macromolecular complexes in situ and resolving their structures at subnanometre resolution [Tegunov et al. (2021)View full citation, Nat. Methods, 18, 186–193]. Despite improvements in data quality as a result of advances in detector technology, microscope stability and stage precision, the analysis and interpretation of tomograms remains challenging due to a low signal-to-noise ratio and reconstruction artifacts stemming from experimental constraints in specimen tilt during data collection resulting in a missing wedge in the Fourier space. Recently, self-supervised deep-learning methods have been proposed for contrast enhancement and reduction of resolution anisotropy in reconstructed tomograms. Here, we evaluate several state-of-the-art deep-learning methods which aim to improve the interpretability of cryo-ET reconstructions, with a focus on their performance on downstream tasks of template matching, sub­tomogram averaging and segmentation. We propose new training architectures and a loss function based on Fourier shell correlation that show improved performance over the standard U-Net with L1/L2 losses. We demonstrate our analysis on four diverse experimental datasets: purified 80S ribosomes, in situ Chlamydomonas reinhardtii, immature HIV-1 virus-like particles and INS-1E cells.

contrast enhancement↗
DOE OSTI2020

What can we learn from N 2 O isotope data? – Analytics, processes and modelling

The isotopic composition of nitrous oxide (N 2 O) provides useful information for evaluating N 2 O sources and budgets. Due to the co–occurrence of multiple N 2 O transformation pathways, it is, however, challenging to use isotopic information to quantify the contribution of distinct processes across variable spatiotemporal scales. Here, we present an overview of recent progress in N 2 O isotopic studies and provide suggestions for future research, mainly focusing on: analytical techniques; production and consumption processes; and interpretation and modelling approaches. Comparing isotope–ratio mass spectrometry (IRMS) with laser absorption spectroscopy (LAS), we conclude that IRMS is a precise technique for laboratory analysis of N 2 O isotopes, while LAS is more suitable for in situ/inline studies and offers advantages for site–specific analyses. When reviewing the link between the N 2 O isotopic composition and underlying mechanisms/processes, we find that, at the molecular scale, the specific enzymes and mechanisms involved determine isotopic fractionation effects. In contrast, at plot–to–global scales, mixing of N 2 O derived from different processes and their isotopic variability must be considered. We also find that dual isotope plots are effective for semi–quantitative attribution of co–occurring N 2 O production and reduction processes. More recently, process–based N 2 O isotopic models have been developed for natural abundance and 15 N–tracing studies, and have been shown to be effective, particularly for data with adequate temporal resolution. Here, despite the significant progress made over the last decade, there is still great need and potential for future work, including development of analytical techniques, reference materials and inter–laboratory comparisons, further exploration of N 2 O formation and destruction mechanisms, more observations across scales, and design and validation of interpretation and modelling approaches. Synthesizing all these efforts, we are confident that the N 2 O isotope community will continue to advance our understanding of N 2 O transformation processes in all spheres of the Earth, and in turn to gain improved constraints on regional and global budgets.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2013

Parallel reduction to hessenberg form with algorithm-based fault tolerance

This paper studies the resilience of a two-sided factorization and presents a generic algorithm-based approach capable of making two-sided factorizations resilient. We establish the theoretical proof of the correctness and the numerical stability of the approach in the context of a Hessenberg Reduction (HR) and present the scalability and performance results of a practical implementation. Our method is a hybrid algorithm combining an Algorithm Based Fault Tolerance (ABFT) technique with diskless checkpointing to fully protect the data. We protect the trailing and the initial part of the matrix with checksums, and protect finished panels in the panel scope with diskless checkpoints. Compared with the original HR (the ScaLAPACK PDGEHRD routine) our fault-tolerant algorithm introduces very little overhead, and maintains the same level of scalability. We prove that the overhead shows a decreasing trend as the size of the matrix or the size of the process grid increases.

Computer Science↗
DOE OSTI2020

Dimensionality Reduction of SDSS Spectra with Variational Autoencoders

High-resolution galaxy spectra contain much information about galactic physics, but the high dimensionality of these spectra makes it difficult to fully utilize the information they contain. We apply variational autoencoders (VAEs), a nonlinear dimensionality reduction technique, to a sample of spectra from the Sloan Digital Sky Survey (SDSS). In contrast to principal component analysis (PCA), a widely used technique, VAEs can capture nonlinear relationships between latent parameters and the data. We find that a VAE can reconstruct the SDSS spectra well with only six latent parameters, outperforming PCA with the same number of components. Different galaxy classes are naturally separated in this latent space, without class labels having been given to the VAE. The VAE latent space is interpretable because the VAE can be used to make synthetic spectra at any point in latent space. For example, making synthetic spectra along tracks in latent space yields sequences of realistic spectra that interpolate between two different types of galaxies. Using the latent space to find outliers may yield interesting spectra: in our small sample, we immediately find unusual data artifacts and stars misclassified as galaxies. In this exploratory work, we show that VAEs create compact, interpretable latent spaces that capture nonlinear features of the data. While a VAE takes substantial time to train (≈1 day for 48,000 spectra), once trained, VAEs can enable the fast exploration of large astronomical data sets.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2012

ISOBAR Preconditioner for Effective and High-throughput Lossless Data Compression

Efficient handling of large volumes of data is a necessity for exascale scientific applications and database systems. To address the growing imbalance between the amount of available storage and the amount of data being produced by high speed (FLOPS) processors on the system, data must be compressed to reduce the total amount of data placed on the file systems. General-purpose loss less compression frameworks, such as zlib and bzlib2, are commonly used on datasets requiring loss less compression. Quite often, however, many scientific data sets compress poorly, referred to as hard-to-compress datasets, due to the negative impact of highly entropic content represented within the data. An important problem in better loss less data compression is to identify the hard-to-compress information and subsequently optimize the compression techniques at the byte-level. To address this challenge, we introduce the In-Situ Orthogonal Byte Aggregate Reduction Compression (ISOBAR-compress) methodology as a preconditioner of loss less compression to identify and optimize the compression efficiency and throughput of hard-to-compress datasets.

Computer Science↗
DOE OSTI2013

Analysis of signals under compositional noise with applications to SONAR data

In this paper, we consider the problem of denoising and classification of SONAR signals observed under compositional noise, i.e., they have been warped randomly along the x-axis. The traditional techniques do not account for such noise and, consequently, cannot provide a robust classification of signals. We apply a recent framework that: 1) uses a distance-based objective function for data alignment and noise reduction; and 2) leads to warping-invariant distances between signals for robust clustering and classification. We use this framework to introduce two distances that can be used for signal classification: a) a y-distance, which is the distance between the aligned signals; and b) an x-distance that measures the amount of warping needed to align the signals. We focus on the task of clustering and classifying objects, using acoustic spectrum (acoustic color), which is complicated by the uncertainties in aspect angles at data collections. Small changes in the aspect angles corrupt signals in a way that amounts to compositional noise. As a result, we demonstrate the use of the developed metrics in classification of acoustic color data and highlight improvements in signal classification over current methods.

47 OTHER INSTRUMENTATION↗
DOE OSTI2016

SU-G-TeP1-11: Predictors of Cardiac and Lung Dose Sparing in DIBH for Left Breast Treatment

Purpose: This retrospective study of left sided whole breast radiation therapy (RT) patients investigates possible predictive parameters correlating to cardiac and left lung dose sparing by deep inspiration breath-hold (DIBH) technique compared to free-breathing (FB). Methods: Thirty-one patients having both DIBH and FB CT scans were included in the study. All patients were planned with a standard step-and-shoot tangential technique using MV photons, with prescription of 50Gy or 50.4Gy. The displacement of the breath hold sternal mark during DIBH, the cardiac contact distances of the axial (CCDax) and parasagittal (CCDps) planes, and lateral-heart-to-chest (LHC) distance on FB CT scans were measured. Lung volumes, mean dose and dose-volume histograms (V5, V10 and V20) were analyzed and compared for heart and left lung for both FB and DIBH techniques. Correlation analysis was performed to identify the predictors for heart and left lung dose sparing. Two-tailed Student’s t-test and linear regression were used for data analysis with significance level of P≤0.05. Results: All dosimetric metrics for the heart and left lung were significantly reduced (P<0.01) with DIBH. Breath hold sternal mark displacement ranged from 0.4–1.8 cm and correlated with mean (P=0.05) and V5 (P=0.02) of heart dose reduction by DIBH. FB lung volume showed correlation with mean lung dose reduction by DIBH (P<0.01). The FB-CCDps and FB-LHC distance had strong positive and negative correlation with FB mean heart dose (P<0.01) and mean heart dose reduction by DIBH (P<0.01), respectively. FB-CCDax showed no correlation with dosimetric changes. Conclusion: DIBH technique has been shown to reduce dose to the heart and left lung. In this patient cohort, FB-CCDps, FB-LHC distance, and FB lung volume served as significant predictors for heart and left lung. These parameters can be further investigated to be used as a tool to better select patients who will benefit from DIBH.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

Investigations of Beam Dynamics Issues at Current and Future Hadron Accelerators (Final Report)

There is a synergy between the fields of Beam Dynamics (BD) in modern particle accelerators and Applied Mathematics (AMa). We have formulated significant problems in BD and have developed and applied tools within the contexts of dynamical systems, topological methods, numerical analysis and scientific computing, probability and stochastic processes, and mathematical statistics. We summarize the three main areas of our AMa work since 2011. First, we continued our study of Vlasov-Maxwell systems. Previously, we developed a state of the art algorithm and code (VM3@A) to calculate coherent synchrotron radiation in single pass systems. In this cycle we carefully analyzed the major expense, namely the integral-over-history (IOH), and developed two approaches to speed up integration. The first strategy uses a representation of the Bessel function J0 in terms of exponentials. The second relies on “local sequences” developed recently for radiation boundary conditions, which are used to reduce computational domains. Although motivated by practicality, both strategies involve interesting and rather deep analysis and approximation theory. As an alternative to VM3@A, we are integrating Maxwell’s equations by a time-stepping method, bypass- ing the IOH, using a Discontinuous Galerkin (DG) method. DG is a generalization of Finite Element and Finite Volume methods. It is spectrally convergent, unlike the commonly used Finite Difference methods, and can handle complicated vacuum chamber geometries. We have applied this in several contexts and have obtained very nice results including an explanation of an experiment at the Canadian Light Source, where the geometry is quite complex. Second, we continued our study of spin dynamics in storage rings. There is much current and proposed activity where spin polarized beams are being used in testing the Standard Model and its modifications. Our work has focused on invariant spin fields (ISFs) and amplitude dependent spin tunes (ADSTs), which are essential for estimating beam polarization. Several algorithms have been developed since the 1980s for computing the ISF, among them the Heinemann- Hoffstaetter method of stroboscopic averaging (SA) which is implemented in the code SPRINT. SA, which computes the ISF by using spin tracking data, can find the ISF to computer precision if it exists and thus can give evidence for existence of the ISF. Central to our work is resolving the ISF conjecture, which says that, off orbital resonance, an ISF exists. Thus Heinemann developed, in his 2010 PhD thesis, a new framework which unifies and generalizes the notions of ISF and ADST by using bundle theory. This lead to a long paper which was a major collaborative effort during the recent cycle. In a nutshell, our bundle approach elegantly unifies the dynamics of spin-1/2 and spin-1 particles, e.g., protons and deuterons. In fact it is well known that these two kinds of dynamics are driven by the same spin transfer matrix and in our approach one sees the deeper reason for this: the spin transfer matrix carries the natural dynamics of a principal bundle whereas the difference between the spin-1/2 and spin-1 dynamics lies in their different geometrical situation, i.e., different underlying associated bundles. Thus one arrives at new results for polarized beams, among them the Invariant Reduction Theorem (IRT) and the Cross Section Theorem (CST). The IRT gives a necessary and sufficient condition for the ISF to exist. The SA technique revealed, 20 years ago, that each ISF can be viewed as a complex agglomerate of spin tracking data. The bundle approach goes one step further by using the IRT and the CST to glue spin tracking data into agglomerates which are candidates for ISFs. We gain insight because the “good” agglomerates, in the presence of an ISF, are very different from the “bad” ones. Finally we mention that the bundle approach has analogies to the approach used in geometrical Yang-Mills theory. Third, we studied X-Ray Free Electron Lasers (FELs), which are electron accelerators producing coherent undulator radiation over a wide range of frequencies from microwaves to x-rays. The photon beams produced in FEL undulators are used to study material samples in biology, material science etc. We developed a mathematical analysis, based on the 6D Lorentz system, of energetic electrons moving through a planar undulator excited by a Maxwell traveling wave field of wavelength λ. Our Method of Averaging perturbation analysis yields non-resonant and near-to-resonant normal form approximations as a function of λ, which we present in two averaging theorems. We prove the theorems in detail, error bounds and giving a tutorial on mathematically rigorous perturbation theory in a context where proofs are easily understood. To our knowledge the planar problem has not been analyzed with the generality here nor has the standard FEL pendulum system, which appears on resonance, been derived with error bounds. In addition to the domains of validity of the normal forms we obtain new insights that require further study, including a more general low gain theory. With a firm foundation in the non-collective case above we are studying the 3D collective case from start up from noise to high gain and saturation. We have formulated the noise start up as a problem of going from the microscopic Klimontovich-Maxwell to the macroscopic Vlasov-Maxwell with a Vlasov correction term. In the 1D setting, we seek an alternative to the phenomenological Vlasov approach which models shot noise by a perturbation on an initial “smooth” density. The 1D wave equation with a Klimontovich source is often the starting point for the 1D FEL high gain theory. We have a new representation of solutions which may lead to new insights.

43 PARTICLE ACCELERATORS↗