Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “sparse data 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 37 records · Page 2
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Characterizing and Mitigating Intraday Variability: Reconstructing Source Structure in Accreting Black Holes with mm-VLBI

The extraordinary physical resolution afforded by the Event Horizon Telescope has opened a window onto the astrophysical phenomena unfolding on horizon scales in two known black holes, M87* and Sgr A*. However, with this leap in resolution has come a new set of practical complications. Sgr A* exhibits intraday variability that violates the assumptions underlying Earth aperture synthesis, limiting traditional image reconstruction methods to short timescales and data sets with very sparse (u, v) coverage. We present a new set of tools to detect and mitigate this variability. We develop a data-driven, model-agnostic procedure to detect and characterize the spatial structure of intraday variability. This method is calibrated against a large set of mock data sets, producing an empirical estimator of the spatial power spectrum of the brightness fluctuations. We present a novel Bayesian noise modeling algorithm that simultaneously reconstructs an average image and statistical measure of the fluctuations about it using a parameterized form for the excess variance in the complex visibilities not otherwise explained by the statistical errors. These methods are validated using a variety of simulated data, including general relativistic magnetohydrodynamic simulations appropriate for Sgr A* and M87*. We find that the reconstructed source structure and variability are robust to changes in the underlying image model. We apply these methods to the 2017 EHT observations of M87*, finding evidence for variability across the EHT observing campaign. The variability mitigation strategies presented are widely applicable to very long baseline interferometry observations of variable sources generally, for which they provide a data-informed averaging procedure and natural characterization of inter-epoch image consistency.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2020

Quantifying uncertainties on fission fragment mass yields with mixture density networks

Probabilistic machine learning techniques can learn both complex relations between input features and output quantities of interest as well as take into account stochasticity or uncertainty within a data set. In this initial work, we explore the use of one such probabilistic network, the mixture density network (MDN), to reproduce fission yields and their uncertainties. In this paper, we study mass yields for the spontaneous fission of 252 Cf, exploring the number of training samples needed for converged predictions, how different levels of uncertainty propagate from the training set to the MDN predictions, and how well physical constraints of the yields—such as normalization and symmetry—are upheld by the algorithm. Finally, we test the ability of the MDN to interpolate between and extrapolate beyond samples in the training set using energy-dependent mass yields for the neutron-induced fission on 235 U. The MDN provides a reliable way to include and predict uncertainties and is a promising path forward for supplementing sparse sets of nuclear data.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2026

Machine Learning Eliminates Reanalysis Warm Bias and Reveals Weaker Winter Surface Cooling Over Arctic Sea Ice

The surface energy budget governs Arctic sea-ice growth/melt, yet observations are sparse, and reanalysis data sets suffer from systematic biases. Here, we train a neural network with observational data to bias-correct hourly ERA5 fluxes over Arctic ice-covered regions (≥70°N; sea-ice concentration >80%) for 1994–2024. Training data cover two full seasonal cycles and different sea-ice regimes. The neural network reduces RMSE for net shortwave radiation by ∼40%, downward longwave radiation by ∼16% and the total surface energy budget by ∼55%, eliminating the wintertime warm bias of ∼4 K in ERA5. Wintertime surface cooling is reduced by ∼50%, yielding thermodynamic ice-growth estimates of ∼80–120 cm, consistent with SMOS–CryoSat satellite thickness increases and in contrast to the 150–200 cm growth implied by ERA5. Our bias-corrected data capture the observed clear/cloudy states of the winter boundary layer and can be used to study Arctic climatology, evaluate climate models and drive sea-ice-ocean models.

Hossain, Akil [Alfred Wegener Institute for Polar ↗
DOE OSTI2025

Exploring the Whole Set of Accurate Sparse Interpretable Models

In data science applications, there are often many models that fit the data well. This phenomenon was called the Rashomon Effect by Leo Breiman. The set of good models is called the Rashomon Set, and the goal of this project is to locate, store, and study the Rashomon sets for classes of interpretable models, including decision trees and generalized additive models.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Transferring a Molecular Foundation Model for Polymer Property Predictions

Transformer-based large language models have remarkable potential to accelerate design optimization for applications such as drug development and material discovery. Self-supervised pretraining of transformer models requires large-scale data sets, which are often sparsely populated in topical areas such as polymer science. Further, state-of-the-art approaches for polymers conduct data augmentation to generate additional samples but unavoidably incur extra computational costs. In contrast, large-scale open-source data sets are available for small molecules and provide a potential solution to data scarcity through transfer learning. In this work, we show that using transformers pretrained on small molecules and fine-tuned on polymer properties achieves comparable accuracy to those trained on augmented polymer data sets for a series of benchmark prediction tasks.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2019

Sparse Recovery for Scientific Data

We have built build and analyzed fundamental mathematical tools that enable sparse recovery of scientific data for high performance computing, experimental facilities and sensor measurement. The proposed work will impact many areas of research across the DOE. We have designed, developed and analyzed fast scalable algorithms for the sparse recovery of scientific data sets that may be extremely large, generated very quickly and distributed across multiple storage devices. The algorithms we developed will be used to process the copious amounts of scientific data generated by the DOE supercomputer simulations and by DOE experimental and observational facilities. They leverage the fact that while the amount of data produced can be extremely large, the actual quantities of interest (QoI's) are often sparse, either in space time, or in some feature space. Mathematically, notions of compressive sensing, including l1 regularized and total variation regularized optimization have played an important part in this effort.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2018

High Resolution Measurements and Modeling in the Arctic

This report details the activity of the project, "High Resolution Modeling and Measurements in the Arctic" spanning Fiscal Years 2016 - 2018 supported by the Sandia National Laboratories Laboratory Directed Research and Development (LDRD) program. The project's primary goal was to test the hypothesis that global climate model bias of low boundary layer clouds lacking liquid water in the Arctic could be improved by increasing horizontal resolution in the model. As model resolution is constrained by computational resources, four different model types were explored and compared to test the project's primary theory. Given the Arctic is a data-sparse region lacking robust data sets of liquid water in clouds, this project also obtained in situ measurements of low clouds with sensors on a tethered balloon system to constrain and compare with the models. Although other model biases remained, the liquid water path generally increased with resolution, supporting the original hypothesis.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2019

Convolutional Dictionary Regularizers for Tomographic Inversion

There has been a growing interest in the use of data-driven regularizers to solve inverse problems associated with computational imaging systems. The convolutional sparse representation model has recently gained attention, driven by the development of fast algorithms for solving the dictionary learning and sparse coding problems for sufficiently large images and data sets. Nevertheless, this model has seen very limited application to tomographic reconstruction problems. In this paper, we present a model-based tomographic reconstruction algorithm using a learnt convolutional dictionary as a regularizer. The key contribution is the use of a data-dependent weighting scheme for the l 1 regularization to construct an effective denoising method that is integrated into the inversion using the Plug-and-Play reconstruction framework. Using simulated data sets we demonstrate that our approach can improve performance over traditional regularizers based on a Markov random field model and a patch-based sparse representation model for sparse and limited-view tomographic data sets.

Singanallur Vaidyanathan, Venkatakrishnan↗
DOE OSTI2021

Characterization of the molten salt FMgNaK through ab initio molecular dynamics and experimental density measurements

The selection of a salt species for use in a molten salt reactor MSR is a key part of any MSR design. However, many salts have sparse or no thermophysical property data sets, especially those with higher melting points. One such salt is the eutectic mixture of (NaF)0.345(KF)0.59(MgF 2 )0.065, or FMgNaK, and is explored here through ab initio molecular dynamics (AIMD) simulation (1023-1273K) as well as experimental measurement of the liquid density using the Archimedean method (973-1223K). Predicted densities from AIMD are similar to experimentally-measured densities, though with more noise, suggesting that the small scale necessitated by AIMD simulations may be problematic for simulating FMgNaK in particular. Here the coefficient of thermal expansion is predicted from simulation, and salt structure is characterized. Mg-F-Mg chaining is observed in the salt network, though the low concentration of Mg inhibits chaining on the scale that is observed in (LiF) 0.67 (BeF 2 ) 0.33 , or FLiBe.

36 MATERIALS SCIENCE↗
DOE OSTI2023

Understanding and Improving Energy Efficiency of Regional Mobility Systems Leveraging System-Level Data

Increased congestion required urban Americans to travel 6.8 billion hours more and purchase 3.1 billion gallons of fuel for a congestion cost of $\$$153 billion, according to the 2019 Urban Mobility Report. How to effectively manage the regional mobility system and improve the energy efficiency presents a big challenge to public agencies. Recent years have witnessed massive multi-jurisdictional multi-modal system-level data from various sources, which provides an unprecedented opportunity to improve the mobility system and its energy efficiency. However, implications of system-level data for mobility and energy efficiency are unclear. Those system-level data sets are siloed, spatially and temporally sparse, biased, not unified, and lacking of insights for system management. Consequently, there is a real need to acquire, fuse, mine and learn from multi-source system-level data to prepare public agencies to deal more effectively with large-scale energy efficiency modeling, management and planning. This project proposes to intensively review inexpensive, replicable and openly-accessible data from multi-modal systems, develop a data-driven system-level modeling framework enabled and validated by data, identify the energy inefficiencies of mobility systems from infrastructure, vehicles, passenger systems, and quantify the benefits of system-level strategies to improve mobility/energy efficiency. In addition, this research develops models to effectively estimate energy consumption and emissions from various types of vehicles on the roadway networks, with high granularity and high fidelity. Traditional models often heavily rely on aggregated infrastructure or vehicle/passenger data, for example, the census survey, land-use, and traffic counts of one or several classes, which may lead to research gaps considering the emerging vehicle technologies. Those models do not contain individual vehicular information. We propose an integrated data-driven method that combines multiple network modeling components, featuring the utilization of state-wide vehicle registration data. The additional vehicle registration data improve the model performance, and produce high-resolution vehicle-specific estimates of emissions and network performance metrics. Two case studies on the Pittsburgh and Philadelphia regional network show that the proposed method can efficiently and effectively estimate the emissions of a large-scale network, and provide valuable information for evaluating common management strategies.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2019

Machine learning framework for assessment of microbial factory performance

Metabolic models can estimate intrinsic product yields for microbial factories, but such frameworks struggle to predict cell performance (including product titer or rate) under suboptimal metabolism and complex bioprocess conditions. On the other hand, machine learning, complementary to metabolic modeling necessitates large amounts of data. Building such a database for metabolic engineering designs requires significant manpower and is prone to human errors and bias. We propose an approach to integrate data-driven methods with genome scale metabolic model for assessment of microbial bio-production (yield, titer and rate). Using engineered E. coli as an example, we manually extracted and curated a data set comprising about 1200 experimentally realized cell factories from ~100 papers. We furthermore augmented the key design features (e.g., genetic modifications and bioprocess variables) extracted from literature with additional features derived from running the genome-scale metabolic model iML1515 simulations with constraints that match the experimental data. Then, data augmentation and ensemble learning (e.g., support vector machines, gradient boosted trees, and neural networks in a stacked regressor model) are employed to alleviate the challenges of sparse, non-standardized, and incomplete data sets, while multiple correspondence analysis/principal component analysis are used to rank influential factors on bio-production. The hybrid framework demonstrates a reasonably high cross-validation accuracy for prediction of E.coli factory performance metrics under presumed bioprocess and pathway conditions (Pearson correlation coefficients between 0.8 and 0.93 on new data not seen by the model).

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2018

A Climate Data Record (CDR) for the global terrestrial water budget: 1984–2010

Closing the terrestrial water budget is necessary to provide consistent estimates of budget components for understanding water resources and changes over time. Given the lack of in situ observations of budget components at anything but local scale, merging information from multiple data sources (e.g., in situ observation, satellite remote sensing, land surface model, and reanalysis) through data assimilation techniques that optimize the estimation of fluxes is a promising approach. Conditioned on the current limited data availability, a systematic method is developed to optimally combine multiple available data sources for precipitation ( P ), evapotranspiration (ET), runoff ( R ), and the total water storage change (TWSC) at 0.5° spatial resolution globally and to obtain water budget closure (i.e., to enforce P -ET- R -TWSC = 0) through a constrained Kalman filter (CKF) data assimilation technique under the assumption that the deviation from the ensemble mean of all data sources for the same budget variable is used as a proxy of the uncertainty in individual water budget variables. The resulting long-term (1984–2010), monthly 0.5° resolution global terrestrial water cycle Climate Data Record (CDR) data set is developed under the auspices of the National Aeronautics and Space Administration (NASA) Earth System Data Records (ESDRs) program. This data set serves to bridge the gap between sparsely gauged regions and the regions with sufficient in situ observations in investigating the temporal and spatial variability in the terrestrial hydrology at multiple scales. The CDR created in this study is validated against in situ measurements like river discharge from the Global Runoff Data Centre (GRDC) and the United States Geological Survey (USGS), and ET from FLUXNET. The data set is shown to be reliable and can serve the scientific community in understanding historical climate variability in water cycle fluxes and stores, benchmarking the current climate, and validating models.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2022

Measurements of the 27 Al(α,n) Thick Target Yield near Threshold

In this work, we present results from direct measurements of the 27 Al(α,n) thick-target yield from laboratory incident energies E α ≈ 3 to 5 MeV, performed with the 3 HeBF 3 Giant Barrel (HeBGB) neutron detector at the Edwards Accelerator Laboratory. Our measurements have a small energy cadence in order to address discrepancies and sparseness of thick-target-yield data sets existing for this energy region. We find general agreement with existing data sets, including yields derived from cross-section data, while resolving a discrepancy between existing thick-target-yield data sets for E α ≈ 4–5 MeV. However, for E α < 3.5 MeV, our results are substantially lower than previous thick-target-yield data and somewhat larger than yields calculated from existing cross-section data. Our data complete the energy range needed for estimates of the 27 Al(α,n) contribution to neutrino and dark matter detector backgrounds and result in increased viability of 27 Al(α,n) as a plasma diagnostic tool at fusion facilities such as the National Ignition Facility.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗
DOE OSTI2017

Gene and transcript abundances of bacterial type III secretion systems from the rumen microbiome are correlated with methane yield in sheep

Background: Ruminants are important contributors to global methane emissions via microbial fermentation in their reticulo-rumens. This study is part of a larger program, characterising the rumen microbiomes of sheep which vary naturally in methane yield (g CH4/kg DM/day) and aims to defne diferences in microbial communities, and in gene and transcript abundances that can explain the animal methane phenotype. Methods: Rumen microbiome metagenomic and metatranscriptomic data were analysed by Gene Set Enrichment, sparse partial least squares regression and the Wilcoxon Rank Sum test to estimate correlations between specifc KEGG bacterial pathways/genes and high methane yield in sheep. KEGG genes enriched in high methane yield sheep were reassembled from raw reads and existing contigs and analysed by MEGAN to predict their phylogenetic origin. Protein coding sequences from Succinivibrio dextrinosolvens strains were analysed using Efective DB to predict bacterial type III secreted proteins. The efect of S. dextrinosolvens strain H5 growth on methane formation by rumen methanogens was explored using co-cultures. Results: Detailed analysis of the rumen microbiomes of high methane yield sheep shows that gene and transcript abundances of bacterial type III secretion system genes are positively correlated with methane yield in sheep. Most of the bacterial type III secretion system genes could not be assigned to a particular bacterial group, but several genes were afliated with the genus Succinivibrio, and searches of bacterial genome sequences found that strains of S. dextrinosolvens were part of a small group of rumen bacteria that encode this type of secretion system. In co-culture experiments, S. dextrinosolvens strain H5 showed a growth-enhancing efect on a methanogen belonging to the order Methanomassiliicoccales, and inhibition of a representative of the Methanobrevibacter gottschalkii clade. Conclusions: This is the frst report of bacterial type III secretion system genes being associated with high methane emissions in ruminants, and identifes these secretions systems as potential new targets for methane mitigation research. The effects of S. dextrinosolvens on the growth of rumen methanogens in co-cultures indicate that bacteriamethanogen interactions are important modulators of methane production in ruminant animals.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2021

Convolutional neural network based non-iterative reconstruction for accelerating neutron tomography *

Abstract Neutron computed tomography (NCT), a 3D non-destructive characterization technique, is carried out at nuclear reactor or spallation neutron source-based user facilities. Because neutrons are not severely attenuated by heavy elements and are sensitive to light elements like hydrogen, neutron radiography and computed tomography offer a complementary contrast to x-ray CT conducted at a synchrotron user facility. However, compared to synchrotron x-ray CT, the acquisition time for an NCT scan can be orders of magnitude higher due to lower source flux, low detector efficiency and the need to collect a large number of projection images for a high-quality reconstruction when using conventional algorithms. As a result of the long scan times for NCT, the number and type of experiments that can be conducted at a user facility is severely restricted. Recently, several deep convolutional neural network (DCNN) based algorithms have been introduced in the context of accelerating CT scans that can enable high quality reconstructions from sparse-view data. In this paper, we introduce DCNN algorithms to obtain high-quality reconstructions from sparse-view and low signal-to-noise ratio NCT data-sets thereby enabling accelerated scans. Our method is based on the supervised learning strategy of training a DCNN to map a low-quality reconstruction from sparse-view data to a higher quality reconstruction. Specifically, we evaluate the performance of two popular DCNN architectures—one based on using patches for training and the other on using the full images for training. We observe that both the DCNN architectures offer improvements in performance over classical multi-layer perceptron as well as conventional CT reconstruction algorithms. Our results illustrate that the DCNN can be a powerful tool to obtain high-quality NCT reconstructions from sparse-view data thereby enabling accelerated NCT scans for increasing user-facility throughput or enabling high-resolution time-resolved NCT scans.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2018

Sparse recovery of undersampled intensity patterns for coherent diffraction imaging at high X-ray energies

Coherent X-ray photons with energies higher than 50 keV offer new possibilities for imaging nanoscale lattice distortions in bulk crystalline materials using Bragg peak phase retrieval methods. However, the compression of reciprocal space at high energies typically results in poorly resolved fringes on an area detector, rendering the diffraction data unsuitable for the three-dimensional reconstruction of compact crystals. To address this problem, we propose a method by which to recover fine fringe detail in the scattered intensity. This recovery is achieved in two steps: multiple undersampled measurements are made by in-plane sub-pixel motion of the area detector, then this data set is passed to a sparsity-based numerical solver that recovers fringe detail suitable for standard Bragg coherent diffraction imaging (BCDI) reconstruction methods of compact single crystals. The key insight of this paper is that sparsity in a BCDI data set can be enforced by recognising that the signal in the detector, though poorly resolved, is band-limited. This requires fewer in-plane detector translations for complete signal recovery, while adhering to information theory limits. Lastly, we use simulated BCDI data sets to demonstrate the approach, outline our sparse recovery strategy, and comment on future opportunities.

36 MATERIALS SCIENCE↗
DOE OSTI2023

Exact Gaussian processes for massive datasets via non-stationary sparsity-discovering kernels

Abstract A Gaussian Process (GP) is a prominent mathematical framework for stochastic function approximation in science and engineering applications. Its success is largely attributed to the GP’s analytical tractability, robustness, and natural inclusion of uncertainty quantification. Unfortunately, the use of exact GPs is prohibitively expensive for large datasets due to their unfavorable numerical complexity of $$O(N^3)$$ O ( N 3 ) in computation and $$O(N^2)$$ O ( N 2 ) in storage. All existing methods addressing this issue utilize some form of approximation—usually considering subsets of the full dataset or finding representative pseudo-points that render the covariance matrix well-structured and sparse. These approximate methods can lead to inaccuracies in function approximations and often limit the user’s flexibility in designing expressive kernels. Instead of inducing sparsity via data-point geometry and structure, we propose to take advantage of naturally-occurring sparsity by allowing the kernel to discover—instead of induce—sparse structure. The premise of this paper is that the data sets and physical processes modeled by GPs often exhibit natural or implicit sparsities, but commonly-used kernels do not allow us to exploit such sparsity. The core concept of exact, and at the same time sparse GPs relies on kernel definitions that provide enough flexibility to learn and encode not only non-zero but also zero covariances. This principle of ultra-flexible, compactly-supported, and non-stationary kernels, combined with HPC and constrained optimization, lets us scale exact GPs well beyond 5 million data points.

97 MATHEMATICS AND COMPUTING↗