SEARCH · Engineering Papers

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.

SourceResearch connectionFrom yearThrough year

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

At least 901 records · Page 50

All sources NASA NTRS DOE OSTI

DOE OSTI2022

EdgeAI: Machine learning via direct attached accelerator for streaming data processing at high shot rate x-ray free-electron lasers

We present a case for low batch-size inference with the potential for adaptive training of a lean encoder model. We do so in the context of a paradigmatic example of machine learning as applied in data acquisition at high data velocity scientific user facilities such as the Linac Coherent Light Source-II x-ray Free-Electron Laser. We discuss how a low-latency inference model operating at the data acquisition edge can capitalize on the naturally stochastic nature of such sources. We simulate the method of attosecond angular streaking to produce representative results whereby simulated input data reproduce high-resolution ground truth probability distributions. By minimizing the mean-squared error between the decoded output of the latent representation and the ground truth distributions, we ensure that the encoding layers and resulting latent representation maintains full fidelity for any downstream task, be it classification or regression. We present throughput results for data-parallel inference of various batch sizes, some with throughput exceeding 100 k images per second. We also show in situ training below 10 s per epoch for the full encoder–decoder model as would be relevant for streaming and adaptive real-time data production at our nation’s scientific light sources.

97 MATHEMATICS AND COMPUTING↗

NASA NTRS2004

New Design and Improvement of Planetary Gear Trains

The development of new types of planetary and planetary face-gear drives is proposed. The new designs are based on regulating backlash between the gears and modifying the tooth surfaces to improve the design. The goal of this work is to obtain a nearly uniform distribution of load between the planet gears. In addition, a new type of planetary face-gear drive was developed in this project.

Handschuh, Robert↗

DOE OSTI2021

ResNet and CycleGAN for pulse shape discrimination of He-4 detector pulses: Recovering pulses conventional algorithms fail to label unanimously

Pulse shape discrimination (PSD) capable detectors, such as He-4, that respond to neutron and gamma-ray 7 interactions have a threshold deposited energy value below which n/γ discrimination vanishes when using 8 conventional PSD algorithms. Recent attempts in applying supervised learning based artificial neural 9 networks for PSD use the pulses in the separated regions to train the networks so they can be used to classify 10 another set of separated pulses. In doing so, pulses previously indistinguishable are not recovered for 11 classification, which would have increased the number of neutron and gamma-ray pulses that could be used 12 for further analysis. Assuming the reason why conventional PSD algorithms have unseparated regions is 13 because the parameter space of the algorithms fail to capture the intrinsic (but subtle) distinguishing 14 behavior of some of the neutron and gamma-ray pulses, a cycle-consistent generative adversarial network 15 (CycleGAN) was trained to amplify those differences and extract well separated neutron and gamma-ray 16 clusters. Results show that, once the network is trained with pulses from separated and unseparated regions, 17 it was able to transform the pulses in the unseparated region to improve the PSD. Subsequent n/γ 18 classification was performed using deep residual network (ResNet) that takes pulses with 512 data points 19 as an input. Two different ResNets were explored – simple ResNet and modified ResNet which takes 20 segmented pulse inputs in the first layer and the corresponding time axis values in the last hidden layer. 21 The later approach enables the network to extract time correlated pulse features to enhance its ability to 22 capture the pulse behaviors relevant for PSD. Although it achieves slightly lower accuracy, 99.41% versus 23 99.89%, based on simply counting the number of correct n/γ labels assigned, compared to the simple 24 ResNet, the modified ResNets architecture was able to decreases the cross-entropy loss function by half, 25 which implies that the correct n/γ labels assigned are less likely to be accidental. PSD parameter 26 distributions based on n/γ classification by ResNet before and after transforming unseparated pulses using 27 CycleGAN show that by enhancing the separation between neutrons and gamma-rays, the transformation 28 helps improve the performance of classifier networks that are trained using labeled dataset. The 29 enhancement of neutron and gamma-ray separation by the CycleGAN increased the PSD figure of merit 30 (FOM) by up to 70% in some regions. Here, the results show that, if a given detector achieves clear separation 31 between neutron and gamma-ray pulses in any energy region, such neural network approaches can help 32 lower the energy threshold for the separation and increasing the number of neutron and gamma-ray pulses 33 that can be used for further analysis.

4He↗

DOE OSTI2025

A segmented approach to modeling building height: Delineating high-rise and low-rise buildings for enhanced height estimation

Understanding building height is imperative to the overall study of energy efficiency, population distribution, urban morphologies, emergency response, among others. Currently, existing approaches for modeling building height at scale are hindered by two pervasive issues. First, there is no consistent approach to quantify what a high-rise building is at a macro scale, leaving researchers unable to accurately compare results across geographies and domains. Second, high-rise buildings represent a small fraction of the built environment, implying data imbalance challenges that negatively affect current approaches. This is a problem of practical relevance since information on high-rise buildings is important for studies on urban heat islands, population dynamics, and pollution dispersion. Here, we introduce a novel approach to map building height which first identifies two distinct distributions within the built environment, with one being composed of low-rise buildings and one composed of high-rise buildings. We then develop an ensemble scheme where discrete specialist models are trained for each subset of low-rise buildings and high-rise buildings to infer building height from morphology features. For experiments mapping heights of 4.85 million buildings in Japan, we show an increase of 34 % in accuracy within 3m error when compared to the current state-of-the-art when modeling high-rise buildings, which based on KNN experimentation we define as any building > 12m . Our findings show that such an ensemble framework outperforms the current state-of-the-art approaches, which is especially relevant in relation to inferring height for high-rise buildings, a prominent issue of existing approaches for mapping the built environment.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2020

Spatiotemporally dynamic implicit large eddy simulation using machine learning classifiers

In this article, we utilize machine learning to dynamically determine if a point on the computational grid requires implicit numerical dissipation for large eddy simulation (LES). The decision making process is learnt through a priori training on quantities derived from direct numerical simulation (DNS) data. In particular, we compute eddy-viscosities obtained through the coarse-graining of DNS quantities and utilize their projection onto a Gaussian distribution to categorize areas that may require dissipation. If our learning determines that closure is necessary, an upwinded scheme is utilized for computing the non-linear Jacobian. In contrast, if it is determined that closure is unnecessary, a symmetric and second-order accurate energy and enstrophy preserving Arakawa scheme is utilized instead. This results in a closure framework that precludes the specification of any model-form for the small scale contributions of turbulence but deploys an appropriate numerical dissipation from explicit closure driven hypotheses. Here, this methodology is deployed for the Kraichnan turbulence test-case and assessed through various statistical quantities such as angle-averaged kinetic energy spectra and vorticity structure functions. Our framework thus establishes a link between the use of explicit LES ideologies for closure and numerical dissipation-based modeling of turbulence leading to improved statistical fidelity of a posteriori simulations.

42 ENGINEERING↗

DOE OSTI2024

Node-degree aware edge sampling mitigates inflated classification performance in biomedical random walk-based graph representation learning

Motivation: Graph representation learning is a family of related approaches that learn low-dimensional vector representations of nodes and other graph elements called embeddings. Embeddings approximate characteristics of the graph and can be used for a variety of machine-learning tasks such as novel edge prediction. For many biomedical applications, partial knowledge exists about positive edges that represent relationships between pairs of entities, but little to no knowledge is available about negative edges that represent the explicit lack of a relationship between two nodes. For this reason, classification procedures are forced to assume that the vast majority of unlabeled edges are negative. Existing approaches to sampling negative edges for training and evaluating classifiers do so by uniformly sampling pairs of nodes. Results: We show here that this sampling strategy typically leads to sets of positive and negative examples with imbalanced node degree distributions. Using representative heterogeneous biomedical knowledge graph and random walk-based graph machine learning, we show that this strategy substantially impacts classification performance. If users of graph machine-learning models apply the models to prioritize examples that are drawn from approximately the same distribution as the positive examples are, then performance of models as estimated in the validation phase may be artificially inflated. We present a degree-aware node sampling approach that mitigates this effect and is simple to implement. Availability and implementation: Our code and data are publicly available at https://github.com/monarch-initiative/negativeExampleSelection.

59 BASIC BIOLOGICAL SCIENCES↗

DOE OSTI2024

Convolutional Variational Autoencoder-based Unsupervised Learning for Power Systems Faults

Classification of power system event data is a growing need, particularly where non-protective relaying-based sensors are used to monitor grid performance. Given the high burden of obtaining event data with appropriate labeling, an unsupervised approach is highly valuable. This approach enables using event data without labeling, which is far easier to obtain. This paper presents an unsupervised learning method to classify and label transients observed in the distribution grid. A Convolutional Variational Autoencoder (CVAE) was developed for this purpose. We demonstrate the efficacy of our approach using the transient data generated from the simulations. The simulation data is used to train the CVAE that identifies different faults as different clusters in the latent space. The clusters are then used as the foundation model to categorize the real-world data.

Alam, Maksudul↗

DOE OSTI2025

Simplified Model and Approach to Transform Infrared Surface Temperature to Film Effectiveness in a Conjugate Heat Transfer Experiment

In the pursuit of more efficient gas turbines, film cooling is a critical technology. This article describes a simplified engineering model based on a one-dimensional thermal resistance network. The model is used to relate film-cooling effectiveness and heat transfer augmentation to local overall cooling effectiveness in a conjugate flat plate experiment. Here, this article presents experimental proof-of-concept data to demonstrate the potential for this model. In contrast to previous approaches, neither the wall heat flux nor the adiabatic wall temperature is required to estimate the local film-cooling performance parameters. The model predicts surface temperatures that are within the experimental uncertainties over the range for which the model is trained and to within five percent when the model is extrapolated to higher coolant channel Reynolds numbers. This article is relevant to conjugate test rigs that can measure the hot-surface temperature distribution with and without film cooling. This information may also be relevant to designers as a method to approximate surface temperatures or used as an approximate heat transfer model for optimization studies.

advanced turbines↗

DOE OSTI2025

Deploying Adversarial Attacks in Super-Resolution Models

Reliable super-resolution methods are crucial for applications like remote sensing, grid resilience and disaster impact analysis, and standoff biometrics. These methods infuse additional high-frequency information into reconstructions, allowing for better contextualization and image intelligence. However, super-resolution models can also introduce hallucinations or other unseen vulnerabilities that could be exploited by an adversary. This is further compounded by the prominence of deep learning in these models, as models are often blindly applied on out-of-distribution images. In this work, we implement adversarial attacks in common open-source super-resolution models and examine their impact on reconstructions and downstream classification tasks. We find that an adversarially trained super-resolution model can produce high-quality reconstructions that degrade downstream classifications. Moreover, these attacks do not require access to low-resolution imagery or class labels at inference time. These results demonstrate the vulnerability of super-resolution methods to malicious actors and motivates the development of a detector for super-resolution adversarial attacks. Further exploration of adversarial attacks in this domain is required to ensure trustworthiness and robustness of super-resolution models for national security applications.

45 MILITARY TECHNOLOGY, WEAPONRY, AND NATIONAL DEF↗

NASA NTRS1991

Optical/IR from ground

Optical/infrared (O/IR) astronomy in the 1990's is reviewed. The following subject areas are included: research environment; science opportunities; technical development of the 1980's and opportunities for the 1990's; and ground-based O/IR astronomy outside the U.S. Recommendations are presented for: (1) large scale programs (Priority 1: a coordinated program for large O/IR telescopes); (2) medium scale programs (Priority 1: a coordinated program for high angular resolution; Priority 2: a new generation of 4-m class telescopes); (3) small scale programs (Priority 1: near-IR and optical all-sky surveys; Priority 2: a National Astrometric Facility); and (4) infrastructure issues (develop, purchase, and distribute optical CCDs and infrared arrays; a program to support large optics technology; a new generation of large filled aperture telescopes; a program to archive and disseminate astronomical databases; and a program for training new instrumentalists)

Strom, Stephen↗

NASA NTRS1991

Use of simulated neural networks of aerial image classification

The utility of one layer neural network in aerial image classification is examined. The network was trained with the delta rule. This method was shown to be useful as a classifier in aerial images with good resolution. It is fast, it is easy to implement, because it is distribution-free, nothing about statistical distribution of the data is needed, and it is very efficient as a boundary detector.

Medina, Frances I.↗

NASA NTRS2008

Automatic Rock Detection and Mapping from HiRISE Imagery

This system includes a C-code software program and a set of MATLAB software tools for statistical analysis and rock distribution mapping. The major functions include rock detection and rock detection validation. The rock detection code has been evolved into a production tool that can be used by engineers and geologists with minor training.

Huertas, Andres↗

NASA NTRS2009

Enhancing Team Performance for Long-Duration Space Missions

Success of exploration missions will depend on skilled performance by a distributed team that includes both the astronauts in space and Mission Control personnel. Coordinated and collaborative teamwork will be required to cope with challenging complex problems in a hostile environment. While thorough preflight training and procedures will equip creW'S to address technical problems that can be anticipated, preparing them to solve novel problems is much more challenging. This presentation will review components of effective team performance, challenges to effective teamwork, and strategies for ensuring effective team performance. Teamwork skills essential for successful team performance include the behaviors involved in developing shared mental models, team situation awareness, collaborative decision making, adaptive coordination behaviors, effective team communication, and team cohesion. Challenges to teamwork include both chronic and acute stressors. Chronic stressors are associated with the isolated and confined environment and include monotony, noise, temperatures, weightlessness, poor sleep and circadian disruptions. Acute stressors include high workload, time pressure, imminent danger, and specific task-related stressors. Of particular concern are social and organizational stressors that can disrupt individual resilience and effective mission performance. Effective team performance can be developed by training teamwork skills, techniques for coping with team conflict, intracrew and intercrew communication, and working in a multicultural team; leadership and teamwork skills can be fostered through outdoor survival training exercises. The presentation will conclude with an evaluation of the special requirements associated with preparing crews to function autonomously in long-duration missions.

Orasanu, Judith M.↗

NASA NTRS2012

OPALS: A COTS-based Tech Demo of Optical Communications

I. Objective: Deliver video from ISS to optical ground terminal via an optical communications link. a) JPL Phaeton/Early Career Hire (ECH) training project. b) Implemented as Class-D payload. c) Downlink at approx.30Mb/s. II. Flight System a) Optical Head Beacon Acquisition Camera. Downlink Transmitter. 2-axis Gimbal. b) Sealed Container Laser Avionics Power distribution Digital I/O board III. Implementation: a) Ground Station - Optical Communications Telescope Laboratory at Table Mountain Facility b) Flight System mounted to ISS FRAM as standard I/F. Attached externally on Express Logistics Carrier.

optical communications↗

NASA NTRS2015

Aerosol Properties From Combined Oxygen A Band Radiances and Lidar

We have developed a new aerosol retrieval technique based on combing high-resolution A band spectra with lidar profiles. Our goal is the development of a technique to retrieve aerosol absorption, one of the critical parameters affecting the global radiation budget and one which is currently poorly constrained by satellite measurements. Our approach relies on two key factors: 1) the use of high spectral resolution (17,000:1) measurements which resolve the A-band line structure, and 2) the use of co-located lidar profile measurements to constrain the vertical distribution of scatterers in the forward model. The algorithm has been developed to be applied to observations from the CALIPSO and OCO-2 satellites, flying in formation as part of the A-train constellation. We describe the approach and present simulated retrievals to illustrate performance potential.

Winker, Dave↗

DOE OSTI2026

Reconstruction of Six-Dimensional Phase Space

A phase space is a mathematical representation of all possible physical states of a system. Particle beams at Fermilab exist within a six-dimensional (6D) phase space defined by three positional components, (x, y, z) and three momentum components, (px, py, pz). To reconstruct this space implies taking measurement data from detectors and mapping out particle behavior using computational methods. The beam detectors, however, are only able to detect spatial distribution among the events of the beam, therefore being limited to positional data. Also, due to the vast number of events in a particle beam, it is extremely difficult to analyze and differentiate every single one’s behavior. However, with Machine Learning (ML), which can distinguish between patterns and map out particle behavior more efficiently. We first used the particle beam software, G4beamline, to simulate a 10,000-event muon beam, adjusting parameters such as initial momentum magnitude (p¬0) and virtual detector position. Using ten virtual detectors, we analyzed p0 values such that minimum 9,990 events were analyzed by every detector. We then input the data from these beam simulations to a C++ program, that randomly selects 100 events, and creates a 2D histogram based on spatial distribution, detector position, and event intensity. This process is repeated 100 times to create 100 histograms per p0 value. These images were then input to a modified ResNet18 Convolutional Neural Network (CNN) for training, and to predict p0 from some unseen set of histograms. The model was accurate when trained on momentum increments of 5 MeV/c and provided with denser training samples around highly variable test values. These results displayed machine learning being able to accurately predict p0 from being trained on different particle behaviors.

Shirlee, Jermain [Fermilab]↗

DOE OSTI2025

Deep Learning without Global Optimization by Random Fourier Neural Networks

Here we introduce a new training algorithm for deep neural networks that utilize random complex exponential activation functions. Our approach employs a Markov chain Monte Carlo sampling procedure to iteratively train network layers, avoiding global and gradient-based optimization while maintaining error control. It consistently attains the theoretical approximation rate for residual networks with complex exponential activation functions, determined by network complexity. Additionally, it enables efficient learning of multiscale and high-frequency features, producing interpretable parameter distributions. Despite using sinusoidal basis functions, we do not observe Gibbs phenomena in approximating discontinuous target functions.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2024

Real Time-Optimal Power Flow-Based Distributed Energy Resource Management System (DERMS)

This project aims to promote lab-proven clean energy technology to commercially scalable versions of the technology, integrate the technology with broader systems, provide extended performance data, and validate the manufacturability and reliability of the technology. The lab-proven technology, RT-OPF DERMS, was developed and validated through previous U.S. Department of Energy-funded efforts, including Advanced Research Projects Agency-Energy funding under the Network Optimized Distributed Energy Systems program and Holy-Cross Energy High Impact Project. In the Advanced Research Projects Agency-Energy Network Optimized Distributed Energy Systems project, the RT-OPF DERMS was developed and implemented in multiple hardware platforms, demonstrating its performance and capabilities in the lab and field environments. The technology was also evaluated and matured via a participation in the U.S. Department of Energy I-Corps program, whose goal is to pair teams of researchers with industry mentors for an intensive 2-month training in which the researchers define technology value propositions, conduct customer discovery interviews, and develop viable market pathways for their technologies. These activities indicate the high technology maturity and Technology Readiness Level of the RT-OPF DERMS.

24 POWER TRANSMISSION AND DISTRIBUTION↗