Search NASA⌕ Search

SEARCH · Search NASA

Results for “Distributed training”

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 271 records · Page 15

Advanced manned space flight simulation and training: An investigation of simulation host computer system concepts

The findings of a preliminary investigation by Southwest Research Institute (SwRI) in simulation host computer concepts is presented. It is designed to aid NASA in evaluating simulation technologies for use in spaceflight training. The focus of the investigation is on the next generation of space simulation systems that will be utilized in training personnel for Space Station Freedom operations. SwRI concludes that NASA should pursue a distributed simulation host computer system architecture for the Space Station Training Facility (SSTF) rather than a centralized mainframe based arrangement. A distributed system offers many advantages and is seen by SwRI as the only architecture that will allow NASA to achieve established functional goals and operational objectives over the life of the Space Station Freedom program. Several distributed, parallel computing systems are available today that offer real-time capabilities for time critical, man-in-the-loop simulation. These systems are flexible in terms of connectivity and configurability, and are easily scaled to meet increasing demands for more computing power.

Montag, Bruce C.↗

Severe Convective Storms in Limited Instability Organized by Pattern and Distribution

Abstract Severe convection occurring in high-shear, low-CAPE (HSLC) environments is a common cool-season threat in the southeastern United States. Previous studies of HSLC convection document the increased operational challenges that these environments present compared to their high-CAPE counterparts, corresponding to higher false-alarm ratios and lower probability of detection for severe watches and warnings. These environments can exhibit rapid destabilization in the hours prior to convection, sometimes associated with the release of potential instability. Here, we use self-organizing maps (SOMs) to objectively identify environmental patterns accompanying HSLC cool-season severe events and associate them with variations in severe weather frequency and distribution. Large-scale patterns exhibit modest variation within the HSLC subclass, featuring strong surface cyclones accompanied by vigorous upper-tropospheric troughs and northward-extending regions of instability, consistent with prior studies. In most patterns, severe weather occurs immediately ahead of a cold front. Other convective ingredients, such as lower-tropospheric vertical wind shear, near-surface equivalent potential temperature (θ e ) advection, and the release of potential instability, varied more significantly across patterns. No single variable used to train SOMs consistently demonstrated differences in the distribution of severe weather occurrence across patterns. Comparison of SOMs based on upper and lower quartiles of severe occurrence demonstrated that the release of potential instability was most consistently associated with higher-impact events in comparison to other convective ingredients. Overall, we find that previously developed HSLC composite parameters reasonably identify high-impact HSLC events. Significance Statement Even when atmospheric instability is not optimal for severe convective storms, in some situations they can still occur, presenting increased challenges to forecasters. These marginal environments may occur at night or during the cool season, when people are less attuned to severe weather threats. Here, we use a sorting algorithm to classify different weather patterns accompanying such storms, and we distinguish which specific patterns and weather system features are most strongly associated with severe storms. Our goals are to increase situational awareness for forecasters and to improve understanding of the processes leading to severe convection in marginal environments.

Meteorology & Atmospheric Sciences↗

Survey of Intelligent Computer-Aided Training

Intelligent Computer-Aided Training (ICAT) systems integrate artificial intelligence and simulation technologies to deliver training for complex, procedural tasks in a distributed, workstation-based environment. Such systems embody both the knowledge of how to perform a task and how to train someone to perform that task. This paper briefly reviews the antecedents of ICAT systems and describes the approach to their creation developed at the NASA Lyndon B. Johnson Space Center. In addition to the general ICAT architecture, specific ICAT applications that have been or are currently under development are discussed. ICAT systems can offer effective solutions to a number of training problems of interest to the aerospace community.

Loftin, R. B.↗

Physics-driven learning of Wasserstein GAN for density reconstruction in dynamic tomography

Object density reconstruction from projections containing scattered radiation and noise is of critical importance in many applications. Existing scatter correction and density reconstruction methods may not provide the high accuracy needed in many applications and can break down in the presence of unmodeled or anomalous scatter and other experimental artifacts. Incorporating machine-learning models could prove beneficial for accurate density reconstruction, particularly in dynamic imaging, where the time evolution of the density fields could be captured by partial differential equations or by learning from hydrodynamics simulations. In this work, we demonstrate the ability of learned deep neural networks to perform artifact removal in noisy density reconstructions, where the noise is imperfectly characterized. Here, we use a Wasserstein generative adversarial network (WGAN), where the generator serves as a denoiser that removes artifacts in densities obtained from traditional reconstruction algorithms. We train the networks from large density time-series datasets, with noise simulated according to parametric random distributions that may mimic noise in experiments. The WGAN is trained with noisy density frames as generator inputs, to match the generator outputs to the distribution of clean densities (time series) from simulations. A supervised loss is also included in the training, which leads to an improved density restoration performance. In addition, we employ physics-based constraints such as mass conservation during the network training and application to further enable highly accurate density reconstructions. Our preliminary numerical results show that the models trained in our frameworks can remove significant portions of unknown noise in density time-series data.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Current and Projected Future Distribution of Plant Community Types for the Southern Seward Peninsula, Alaska

Landscape scale maps of plant community distribution at 5 m resolution were developed for southern Seward Peninsula. A Random Forest based environmental niche model was developed using topographic and climate data sets. Models were trained for the 2010-2019 period and applied to develop plant community distributions for contemporary (2010-2019) period. Trained model was applied to project the plant community distributions under future climate. Using downscaled projections for RCP8.5 scenario from five climate models (CCSM4, GFDL-CM3, GISS-E2-R, IPSL-CM5A-LR, MRI-CGCM3) and multi-model mean, a six member ensemble of plant community distribution were developed for current decade (2010-2019) and for future decades (2020-2029, 2030-2039, 2040-2049, 2050-2059). Contains *.tif and *.sh files. The Next-Generation Ecosystem Experiments: Arctic (NGEE Arctic), was a research effort to reduce uncertainty in Earth System Models by developing a predictive understanding of carbon-rich Arctic ecosystems and feedbacks to climate. NGEE Arctic was supported by the Department of Energy's Office of Biological and Environmental Research. The NGEE Arctic project had two field research sites: 1) located within the Arctic polygonal tundra coastal region on the Barrow Environmental Observatory (BEO) and the North Slope near Utqiagvik (Barrow), Alaska and 2) multiple areas on the discontinuous permafrost region of the Seward Peninsula north of Nome, Alaska. Through observations, experiments, and synthesis with existing datasets, NGEE Arctic provided an enhanced knowledge base for multi-scale modeling and contributed to improved process representation at global pan-Arctic scales within the Department of Energy's Earth system Model (the Energy Exascale Earth System Model, or E3SM), and specifically within the E3SM Land Model component (ELM).

54 ENVIRONMENTAL SCIENCES↗

Three-Dimensional Structure and Modeling of a Normal Bifurcated Shock Train from Experimental Measurements

An asymmetric, normal bifurcated shock train in a Mach 2 constant area, rectangular duct is investigated in the University of Michigan Direct Connect Isolator facility. High-speed schlieren imaging, wall static pressure measurements, surface oil flow visualization, and particle image velocimetry of the shock train are synthesized into a three-dimensional representation. This visualization is then used to inform the underlying flow physics of the distributed fluid dynamical processes along the shock train. The detailed, three-dimensional morphology of the flow profile entering the shock train is shown to have a significant impact on the separated flow morphology of the shock train. This results in a skewed supersonic core flow with alternating separation bubbles in the surrounding boundary layers. Finally, the efficacy of the prevailing pseudo-shock models in the literature is analyzed with the support of the available flow measurements. None are found to model both the pressure and streamwise Mach number profile of the UMDCI pseudo-shock. Note: This presentation is accompanied by the video presentation that can be downloaded for viewing and is formatted as an mp4 with a runtime of 10 min 15 secs.

Hypersonics↗

Neural network modeling of in situ fluid-filled pore size distributions in subsurface shale reservoirs under data constraints

Subsurface nuclear magnetic resonance (NMR) logs acquired in the wellbore environment are sensitive to fluid-filled pore size distribution, fluid mobility, permeability, and porosity in the near-wellbore reservoir volume. NMR response of a formation layer is processed to extract the T2 distribution, which approximates the fluid-filled pore size distribution. NMR logs are acquired in limited number of wells due to financial and operational challenges, which adversely affects reservoir characterization. We developed here two neural-network-based machine learning techniques, long short-term memory (LSTM) network and variational autoencoder with a convolutional layer (VAEc) network, to process the ‘easy-to-acquire’ formation mineral and fluid saturation logs to generate synthetic NMR T2 distributions in the absence of ‘hard-to-acquire’ NMR T2 distribution log. Both the predictive models are trained and tested on limited wireline log measurements randomly selected from a 300-ft depth interval of the Bakken shale formation. Synthesis performances of LSTM and VAEc models in terms of R 2 are 0.78 and 0.75, respectively. Noise is inevitable in logging data due to the complex wellbore and formation conditions. Notably, both the predictive models robustly synthesize the fluid-filled pore size distributions in the presence of 50% noise in input logs and 30% noise in training T2 data. The performance of the proposed methodology improves with access to larger volume of training data from other formation types. The proposed method is critical to the synthesis of in situ fluid-filled pore size distributions in shale formations under data constraints due to financial and operational challenges.

58 GEOSCIENCES↗

Learning generative neural networks with physics knowledge

Deep generative neural networks have enabled modeling complex distributions, but incorporating physics knowledge into the neural networks is still challenging and is at the core of current physics-based machine learning research. To this end, we propose a physics generative neural network (PhysGNN), a new class of generative neural networks for learning unknown distributions in a physical system described by partial differential equations (PDE). PhysGNN couples PDE systems with generative neural networks. It is a fully differentiable model that allows back-propagation of gradients through both numerical PDE solvers and generative neural networks, and is trained by minimizing the discrete Wasserstein distance between generated and observed probability distributions of the PDE outputs using the stochastic gradient descent method. Moreover, PhysGNN does not require adversarial training like standard generative neural networks, which offers better stability than adversarial training. We show that PhysGNN can learn complex distributions in stochastic inverse problems, where conventional methods such as maximum likelihood estimation and momentum matching methods may be inapplicable when little knowledge is known about the form of unknown distributions or the physical model is too complex. Furthermore, our method allows physics-based generative neural network training for learning complex distributions in the context of differential equations.

97 MATHEMATICS AND COMPUTING↗

Multiple hot images from an obscuration in an intense laser beam through cascaded Kerr medium disks

We present a theoretical investigation on the formation of hot images in an intense laser beam through cascaded Kerr medium disks, to disclose the distribution and intensity of hot images in high-power disk amplifiers. It is shown that multiple hot images from an obscuration may be formed, instead of one hot image as reported previously in the literature. This gives a clear explanation for the curious damage pattern of hot images, namely, damage sites appearing on alternating optics in periodic trains. Further analysis demonstrates that the distribution and intensity of hot images depend closely on the number of Kerr medium disks, the distance from the obscuration to the front of the first disk downstream, the space between two neighboring disks, and the thickness and B integral of each disk. Moreover, we take two cascaded Kerr medium disks for example to detail multiple hot images from an obscuration and confirm the theoretical results by numerical simulations.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Biological shield design and analysis of KIPT accelerator-driven subcritical facility.

Argonne National Laboratory of the United States and Kharkov Institute of Physics and Technology of Ukraine have been collaborating on the conceptual design development of an electron accelerator-driven subcritical facility. The facility will be utilized for performing basic and applied nuclear research, producing medical isotopes, and training young nuclear specialists. This paper presents the design and analyses of the biological shield performed for the top section of the facility. The neutron source driving the subcritical assembly is generated from the interaction of a 100-kW electron beam with a natural uranium target. The electron energy is in the range of 100 to 200 MeV, and it has a uniform spatial distribution. The shield design and the associated analyses are presented including different parametric studies. In the analyses, a significant effort was dedicated to the accurate prediction of the radiation dose outside the shield boundary as a function of the shield thickness without geometrical approximations or material homogenization. The MCNPX Monte Carlo code was utilized for the transport calculation of electrons, photons, and neutrons. Weight window variance-reduction techniques were introduced, and the dose equivalent outside the shield can be calculated with reasonably good statistics.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Machine learning approaches for structural and thermodynamic properties of a Lennard-Jones fluid

Predicting the functional properties of many molecular systems relies on understanding how atomistic interactions give rise to macroscale observables. However, current attempts to develop predictive models for the structural and thermodynamic properties of condensed-phase systems often rely on extensive parameter fitting to empirically selected functional forms whose effectiveness is limited to a narrow range of physical conditions. Here, we illustrate how these traditional fitting paradigms can be superseded using machine learning. Specifically, we use the results of molecular dynamics simulations to train machine learning protocols that are able to produce the radial distribution function, pressure, and internal energy of a Lennard-Jones fluid with increased accuracy in comparison to previous theoretical methods. The radial distribution function is determined using a variant of the segmented linear regression with the multivariate function decomposition approach developed by Craven et al. [J. Phys. Chem. Lett. 11, 4372 (2020)]. The pressure and internal energy are determined using expressions containing the learned radial distribution function and also a kernel ridge regression process that is trained directly on thermodynamic properties measured in simulation. The presented results suggest that the structural and thermodynamic properties of fluids may be determined more accurately through machine learning than through human-guided functional forms.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Improving Grid Resilience through Informed Decision-making (IGRID)

The transformation of the distribution grid from a centralized to decentralized architecture, with bi-directional power and data flows, is made possible by a surge in network intelligence and grid automation. While changes are largely beneficial, the interface between grid operator and automated technologies is not well understood, nor are the benefits and risks of automation. Quantifying and understanding the latter is an important facet of grid resilience that needs to be fully investigated. The work described in this document represents the first empirical study aimed at identifying and mitigating the vulnerabilities posed by automation for a grid that for the foreseeable future will remain a human-in-the-loop critical infrastructure. Our scenario-based methodology enabled us to conduct a series of experimental studies to identify causal relationships between grid-operator performance and automated technologies and to collect measurements of human performance as a function of automation. Our findings, though preliminary, suggest there are predictive patterns in the interplay between human operators and automation, patterns that can inform the rollout of distribution automation and the hiring and training of operators, and contribute in multiple and significant ways to the field of grid resilience.

24 POWER TRANSMISSION AND DISTRIBUTION↗

The Fate of Saharan Dust Across the Atlantic: An Integrated Modeling and Observational Study of the TC4 Field Campaign

During the NASA TC-4 field campaign in July 2007, several Saharan dust events were observed over the Caribbean basin. A-Train observations suggest that these Saharan dust events are confined the Caribbean and rarely transported across Central America to the Pacific Ocean. We investigate the nature of this barrier to dust transport using the NASA GEOS-5 atmospheric general circulation model. Our simulations with GEOS-5 are driven by the Modern Era Retrospective-Analysis for Research and Applications (MERRA) meteorological analyses, and include online simulation of aerosol distributions using a version of the Goddard Chemistry, Aerosol, Radiation, and Transport (GOCART) model. Simulated dust distributions are evaluated using A-Train observations from MODIS and CALIOP, as well as MISR and ground-based AERONET sun photometers, and show good agreement with the observations in terms of the timing and magnitude of dust events. A component analysis of the dust transport and removal pathways is used to evaluate the relative roles of these processes in establishing the observed dust transport barrier. From this analysis, we show that while both atmospheric dynamics and wet removal contribute towards the Caribbean dust barrier, northward dust transport is the more dominant term. Additional simulations are performed to ascertain the sensitivity of our results to uncertain loss processes (i.e., wet removal) in our model.

Nowottnick, E.↗

LTAU-FF: Loss Trajectory Analysis for Uncertainty in atomistic Force Fields

Model ensembles are effective tools for estimating prediction uncertainty in deep learning atomistic force fields. However, their widespread adoption is hindered by high computational costs and overconfident error estimates. In this work, we address these challenges by leveraging distributions of per-sample errors obtained during training and employing a distance-based similarity search in the model latent space. Our method, which we call LTAU (Loss Trajectory Analysis for Uncertainty), efficiently estimates the full probability distribution function of errors for any test point using the logged training errors, achieving speeds that are 2–3 orders of magnitudes faster than typical ensemble methods and allowing it to be used for tasks where training or evaluating multiple models would be infeasible. We apply LTAU towards estimating parametric uncertainty in atomistic force fields (LTAU-FF), demonstrating that it produces well-calibrated confidence intervals and predicts errors that correlate strongly with the true errors for data near the training domain. Furthermore, we show that the errors predicted by LTAU-FF can be used in practical applications for detecting out-of-domain data, tuning model performance, and predicting failure during simulations. We believe that LTAU will be a valuable tool for uncertainty quantification in atomistic force fields and is a promising method that should be further explored in other domains of machine learning.

97 MATHEMATICS AND COMPUTING↗

Estimating Watershed Subsurface Permeability From Stream Discharge Data using Deep Neural Networks, Frontiers in Earth Science: Dataset

This data package contains watershed modeling inputs and outputs as well as deep neural networks training and testing results used in "Estimating Watershed Subsurface Permeability From Stream Discharge Data using Deep Neural Networks" (Cromwell et al., 2021). We train various deep neural network models with different architectures to predict subsurface permeability from stream discharge hydrograph at the watershed outlet. The training data are obtained from ensemble simulations of hydrographs corresponding to an permeability ensemble using a fully-distributed, integrated surface-subsurface hydrologic model. The trained model is then applied to estimate the permeability of the real watershed using its observed hydrograph at the outlet. Our study demonstrates that the permeabilities of the soil and geologic facies that make significant contributions to the outlet discharge can be more accurately estimated from the discharge data. Their estimations are also more robust with observation errors. Compared to the traditional ensemble smoother method, DNNs show stronger performance in capturing the nonlinear relationship between permeability and stream hydrograph to accurately estimate permeability. Our study sheds new light on the value of the emerging deep learning methods in assisting integrated watershed modeling by improving parameter estimation, which will eventually reduce the uncertainty in predictive watershed models. For detailed information regarding watershed model description and DNNs setup, please refer to Cromwell et al., 2021. The deep_learning.zip file contains the inputs and outputs for the different deep neural networks. The watershed_model.zip file contains the inputs and outputs for watershed simulation using ATS. The figures.zip file contains raw figures and their corresponding scripts. For a more detailed description, please see the README.md file within each compressed file.

54 ENVIRONMENTAL SCIENCES↗

Secondary Fragmentation of Comet Shoemaker-Levy 9 and the Ramifications for the Progenitor's Breakup in July 1992

Comprehensive analysis of discrete events of secondary fragmentation leads to a conceptually new understanding of the process of disintegration of comet Shoemaker-Levy 9. We submit that the jovian tidal forces inflicted extensive cracks throughout the interior of the original nucleus but did not split it apart. The initial disruption was apparently accomplished by stresses exerted on the cracked object by its fast rotation during the early post-perijove period of time. We argue that this disruption was in fact a rapid sequence of episodes during July 1992 that gave birth to the 12 on-train, or primary, fragments: A, C, D, E, G, H, K, L, Q (later Q(sub 1)), R, S, and W. The discrete events of secondary fragmentation, which gave birth to the off-train fragments, are understood in this scenario as stochastic manifestations of the continuing process of progressive disintegration. Of the 13 off-train fragments considered, nine were secondary--B, F, G(sub 2), M, N, P (later P(sub 2) or P(sub 2a)), Q(sub 2), U, and V--and four tertiary (J, P(sub 1), P(sub 2b), and T). The separation parameters of 11 off-train fragments were determined. The vectorial distribution of separation velocities of these fragments shows a strong concentration toward a great circle, unquestionably an effect of the approximately conserved angular momentum of the progenitor comet since the time of its initial disruption. Also apparent is their clumping (except for P(sub 1)) to a segment along the great circle, implying that the fragments were consistently released from one side of their parents, thus explaining for the first time why the off-train fragments preferentially appeared on one side of the nuclear train. In order to obtain a consistent solution, our model requires that the points of separation be on the antisolar side of the parent fragments, where thermal stresses are likely to enhance the effect of rotation. The episodes of secondary fragmentation are found to have nine months after the close encounter with Jupiter in early July 1992, and the separation velocities ranged between 0.36 and 1.7 m/s. The spin-axis position is determined to have been nearly in the jovicentric orbit plane, which rules out the Asphaug-Benz-Solem strengthless aggregate model as a plausible breakup hypothesis. Since the separation velocities are rotational in nature, they cannot substantially exceed the critical limit for centrifugal breakup and offer an estimate for the original nuclear dimensions. The comet's nucleus is found to have been approximately 10 km in diameter and spinning rapidly. With the exception of P(sub 1) and apparently also P(sub 2) and F, no nongravitational deceleration was detected in the motions of the off-train fragments. Serious doubts are cast on continuing appreciable activity of any of these fragments. Indeed, when it was necessary to introduce a deceleration into the equations of motion, the effect appears to have been due to the action of solar radiation pressure on the centroid of centimeter-sized particulates in the disintegrating condensations.

Sekanina, Z.↗

Microstructure Segmentation with Deep Learning Encoders Pre-Trained on a Large Microscopy Dataset

This study examined the improvement of microscopy segmentation accuracy by transfer learning from a large dataset of microscopy images called MicroNet. Many neural network encoder architectures, including VGG, Inception, and ResNet, were trained on over 100,000 labelled microscopy images from 54 classes. These pre-trained encoders were then embedded into multiple segmentation architectures including U-Net and DeepLabV3+ to evaluate segmentation performance on newly created benchmark microscopy datasets. Compared to ImageNet pre-training, models pre-trained on MicroNet generalized better to out-of-distribution micrographs taken under different imaging and sample conditions and were more accurate with less training data. When training with only a single Ni-superalloy image, pre-training on MicroNet produced a 72.2 percent reduction in relative segmentation error. These results suggest that transfer learning from large in-domain datasets generate models with learned feature representations that are more useful for downstream tasks and will likely improve any microscopy image analysis technique that can leverage pre-trained encoders.

machine learning↗

Microstructure Segmentation With Deep Learning Encoders Pre-Trained on a Large Microscopy Dataset

This study examined the improvement of microscopy segmentation intersection over union accuracy by transfer learning from a large dataset of microscopy images called MicroNet. Many neural network encoder architectures were trained on over 100,000 labeled microscopy images from 54 material classes. These pre-trained encoders were then embedded into multiple segmentation architectures including UNet and DeepLabV3+ to evaluate segmentation performance on created benchmark microscopy datasets. Compared to ImageNet pre-training, models pre-trained on MicroNet generalized better to out-of-distribution micrographs taken under different imaging and sample conditions and were more accurate with less training data. When training with only a single Ni-superalloy image, pre-training on MicroNet produced a 72.2% reduction in relative intersection over union error. These results suggest that transfer learning from large in-domain datasets generate models with learned feature representations that are more useful for downstream tasks and will likely improve any microscopy image analysis technique that can leverage pre-trained encoders.

machine learning↗