SEARCH · Engineering Papers

Results for “DNN”

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 163 records · Page 9

All sources NASA NTRS DOE OSTI

DOE OSTI2020

Identifying the nature of the QCD transition in relativistic collision of heavy nuclei with deep learning

Using deep convolutional neural network (CNN), the nature of the QCD transition can be identified from the final-state pion spectra from hybrid model simulations of heavy-ion collisions that combines a viscous hydrodynamic model with a hadronic cascade “after-burner”. Two different types of equations of state (EoS) of the medium are used in the hydrodynamic evolution. The resulting spectra in transverse momentum and azimuthal angle are used as the input data to train the neural network to distinguish different EoS. Different scenarios for the input data are studied and compared in a systematic way. A clear hierarchy is observed in the prediction accuracy when using the event-by-event, cascade-coarse-grained and event-fine-averaged spectra as input for the network, which are about 80%, 90% and 99%, respectively. A comparison with the prediction performance by deep neural network (DNN) with only the normalized pion transverse momentum spectra is also made. High-level features of pion spectra captured by a carefully-trained neural network were found to be able to distinguish the nature of the QCD transition even in a simulation scenario which is close to the experiments.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2020

Toward Large-Scale Image Segmentation on Summit

Semantic segmentation of images is an important computer vision task that emerges in a variety of application domains such as medical imaging, robotic vision and autonomous vehicles to name a few. While these domain-specific image analysis tasks involve relatively small image sizes (~ 10 2 × 10 2 ), there are many applications that need to train machine learning models on image data with extents that are orders of magnitude larger (~10 4 × 10 4 ). Training deep neural network (DNN) models on large extent images is extremely memory-intensive and often exceeds the memory limitations of a single graphical processing unit, a hardware accelerator of choice for computer vision workloads. Here, an efficient, sample parallel approach to train U-Net models on large extent image data sets is presented. Its advantages and limitations are analyzed and near-linear strong-scaling speedup demonstrated on 256 nodes (1536 GPUs) of the Summit supercomputer. Using a single node of the Summit supercomputer, an early evaluation of a recently released model parallel framework called GPipe is demonstrated to deliver ~ 2X speedup in executing a U-Net model with an order of magnitude larger number of trainable parameters than reported before. Performance bottlenecks for pipelined training of U-Net models are identified and mitigation strategies to improve the speedups are discussed. Together, these results open up the possibility of combining both approaches into a unified scalable pipelined and data parallel algorithm to efficiently train U-Net models with very large receptive fields on data sets of ultra-large extent images.

Seal, Sudip↗

DOE OSTI2023

XploreNAS : Explore Adversarially Robust and Hardware-efficient Neural Architectures for Non-ideal Xbars

Compute In-Memory platforms such as memristive crossbars are gaining focus as they facilitate acceleration of Deep Neural Networks (DNNs) with high area and compute efficiencies. However, the intrinsic non-idealities associated with the analog nature of computing in crossbars limits the performance of the deployed DNNs. Furthermore, DNNs are shown to be vulnerable to adversarial attacks leading to severe security threats in their large-scale deployment. Thus, finding adversarially robust DNN architectures for non-ideal crossbars is critical to the safe and secure deployment of DNNs on the edge. This work proposes a two-phase algorithm-hardware co-optimization approach called XploreNAS that searches for hardware efficient and adversarially robust neural architectures for non-ideal crossbar platforms. We use the one-shot Neural Architecture Search approach to train a large Supernet with crossbar-awareness and sample adversarially robust Subnets therefrom, maintaining competitive hardware efficiency. Our experiments on crossbars with benchmark datasets (SVHN, CIFAR10, CIFAR100) show up to ~8–16% improvement in the adversarial robustness of the searched Subnets against a baseline ResNet-18 model subjected to crossbar-aware adversarial training. We benchmark our robust Subnets for Energy-Delay-Area-Products (EDAPs) using the Neurosim tool and find that with additional hardware efficiency–driven optimizations, the Subnets attain ~1.5–1.6× lower EDAPs than ResNet-18 baseline.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2023

pnnl/MBDRL

Model-based Deep Reinforcement Learning for Real-time Grid Emergency Voltage Control. A model-based Deep Reinforcement Learning (DRL) framework where a deep neural network (DNN)-based surrogate model is utilized within the control policy learning framework, making the learning process faster and more sample efficient

Yin, Tim↗

DOE OSTI2024

Codes for "Shallow Soil Response to a Buried Chemical Explosion with Geophones and Distributed Acoustic Sensing" DAG - 01101646

The codes reproduce the figures of the manuscript entitled "Shallow Soil Response to a Buried Chemical Explosion with Geophones and Distributed Acoustic Sensing" submitted to Journal of Geophysical Research - Solid Earth. Geophone data and Distributed acoustic sensing (DAS) data recorded during the Phase II of the The Source Physics Experiment (SPE) along a fiber-optic cable offshore were processed to understand the response of the shallow subsurface to an explosion. This Ground-based Nuclear Detonation Detection (GNDD), Low Yield Nuclear Monitoring (LYNM), and Source Physics Experiment (SPE) research was funded by the National Nuclear Security Administration, Defense Nuclear Nonproliferation Research and Development (NNSA DNN R&D).

Viens, Loic↗

DOE OSTI2024

UFNet: Joint U-Net and Fully Connected Neural Network to Bias Correct Precipitation Predictions from

Paper information. Shuang Yu, Indrasis Chakraborty, Gemma J. Anderson, Donald D. Lucas, Yannic Lops, and Daniel Galea. UFNet: Joint U-Net and fully connected neural network to bias correct precipitation predictions from climate models. Artificial Intelligence for the Earth Systems, 2024. Overview. This work develops the UFNet methodology to correct E3SM historical precipitation projection bias. The UFNet deep learning framework consists of a two-part architecture: a U-Net convolutional network to capture the spatiotemporal distribution of precipitation and a fully connected network to capture the distribution of higher-order statistics. The joint network, termed UFNet, can simultaneously improve the spatial structure of the modeled precipitation and capture the distribution of extreme precipitation values. Below we provide guidance for applying UFNet to correct the Energy Exascale Earth System Model (E3SM; Golaz et al. 2019) daily precipitation projection over the contiguous United States (CONUS). Getting started 1. Obtain the historical climate simulation and observation data. The E3SM historical simulation data are available through https://aims2.llnl.gov/search/cmip6/. The CPC unified gauge-based analysis of daily precipitation can be found through https://psl.noaa.gov/data/gridded/data.cpc.globalprecip.html. The ECMWF atmospheric reanalysis of the 20th century (ERA-20C) data are available through https://www.ecmwf.int/en/forecasts/datasets/reanalysis-datasets/era-20c. The spatial resolution of E3SM and observed datasets are both regridded to a common 1° resolution grid using conservative interpolation. The regridded E3SM, CPC and ERA-20C with 1° resolution can be found throught ./data/. 2. Train the fully connected network (DNN) Python train_dnn.py 3. Train the UFNet Python train_ufnet.py 4. Evaluation and compared with the baseline Python evaluation.py

Lucas, DonaldD↗

DOE OSTI2022

Deep Neural Network High Spatiotemporal Resolution Precipitation Estimation (Deep-STEP) Using Passive Microwave and Infrared Data

Recent developments in “headline-making” deep neural networks (DNNs), specifically convolutional neural networks (CNNs), along with advancements in computational power, open great opportunities to integrate massive amounts of real-time observations to characterize spatiotemporal structures of surface precipitation. This study aims to develop a CNN algorithm, named Deep Neural Network High Spatiotemporal Resolution Precipitation Estimation (Deep-STEP), that ingests direct satellite passive microwave (PMW) brightness temperatures (Tbs) at emission and scattering frequencies combined with infrared (IR) Tbs from geostationary satellites and surface information to automatically extract geospatial features related to the precipitable clouds. These features allow the end-to-end Deep-STEP algorithm to instantaneously map surface precipitation intensities with a spatial resolution of 4 km. The main advantages of Deep-STEP, as compared to current state-of-the-art techniques, are 1) it learns and estimates complex precipitation systems directly from raw measurements in near–real time, 2) it uses the automatic spatial neighborhood feature extraction approach, and 3) it fuses coarse-resolution PMW footprints with IR images to reliably retrieve surface precipitation at a high spatial resolution. We anticipate our proposed DNN algorithm to be a starting point for more sophisticated and efficient precipitation retrieval systems in terms of accuracy, fine spatial pattern detection skills, and computational costs.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2020

Towards Query-Efficient Black-Box Adversary with Zeroth-Order Natural Gradient Descent

Despite the great achievements of the modern deep neural networks (DNNs), the vulnerability/robustness of state-of-the-art DNNs raises security concerns in many application domains requiring high reliability. Various adversarial attacks are proposed to sabotage the learning performance of DNN models. Among those, the black-box adversarial attack methods have received special attentions owing to their practicality and simplicity. Black-box attacks usually prefer less queries in order to maintain stealthy and low costs. However, most of the current black-box attack methods adopt the first-order gradient descent method, which may come with certain deficiencies such as relatively slow convergence and high sensitivity to hyper-parameter settings. In this paper, we propose a zeroth-order natural gradient descent (ZO-NGD) method to design the adversarial attacks, which incorporates the zeroth-order gradient estimation technique catering to the black-box attack scenario and the second-order natural gradient descent to achieve higher query efficiency. The empirical evaluations on image classification datasets demonstrate that ZO-NGD can obtain significantly lower model query complexities compared with state-of-the-art attack methods.

Zhao, Pu↗

DOE OSTI2026

Data-Driven Modeling and Correction of Vehicle Dynamics

We develop a data-driven framework for learning and correcting nonautonomous vehicle dynamics. Physics-based vehicle models are often simplified for tractability and therefore exhibit inherent model-form uncertainty, motivating the need for data-driven correction. Moreover, nonautonomous dynamics are governed by time-dependent control inputs, which pose challenges in learning predictive models directly from temporal snapshot data. To address these, we reformulate the vehicle dynamics via a local parameterization of the time-dependent inputs, yielding a modified system composed ofa sequence of local parametric dynamical systems. Here, we approximate these parametric systems using two complementary approaches. First, we employ the dimension reduction and interpolation in parameter space (DRIPS) methodology to construct efficient linear surrogate models, equipped with lifted observable spaces and manifold-based operator interpolation. This enables data-efficient learning of vehicle models whose dynamics admit accurate linear representations in the lifted spaces. Second, for more strongly nonlinear systems, we employ flow map learning (FML), a deep neural network (DNN) approach that approximates the parametric evolution map without requiring special treatment of nonlinearities. We further extend FML with a transfer-learning-based model correction procedure, enabling the correction of misspecified prior models using only a sparse set of high-fidelity or experimental measurements, without assuming a prescribed form for the correction term. Through a suite of numerical experiments on unicycle, simplified bicycle, and slip-based bicycle models, we demonstrate that DRIPS offers robust and highly data-efficient learning of nonautonomous vehicle dynamics, while FML provides expressive nonlinear modeling and effective correction of model-form errors under severe data scarcity.

data-driven modeling↗

DOE OSTI2020

Mixed-Material Scintillators (MMSS Quarterly Report FY20Q2)

This project aims to invent, model, and prototype mixed-material scintillator systems (MMSSs), a new class of radiation detectors that use heterogenous internal structures of different scintillating materials to detect additional properties of radiation. These internal structures can be produced using additive manufacture (3D printing) of scintillator, a currently emerging application of additive manufacture technology. MMSSs combine the low cost and complexity of conventional scintillation detectors with capabilities currently only available in more expensive and complex detectors. By identifying promising MMSS designs, this project will enable a new class of detectors to meet DNN’s mission needs for SNM detection. This quarter, we delivered one the MMSS Particle-ID Modeling report, a major deliverable of this project. This report describes our results on simulations of the particle-ID (PID) class of detectors. We found that these detectors outperform competing approaches for neutron-gamma discrimination, neutron source pointing, and neutron spectroscopy. Further highlights of the report are described below. During this quarter, We alerted HQ to a change in schedule for the remaining simulations planned for this project: we plan to put off further simulation and reporting of the position-resolving (PR) detector class until July. This change in schedule has allowed us to go into additional detail in the PID work that is both very promising and motivates our upcoming proposal. The other active task in this project is prototyping 3D-printed scintillators with the aim of realizing the designs we’ve simulated. On that front, we had three focuses: measuring the light output, increasing the size of printed parts, and combining blue and green scintillators. We succeeded in measuring the light output, showing an output of 30% of commercial standard. This result leaves room for improvement but is within striking distance of where we need to be. We were on track to demonstrate a key requirement for increasing the size of parts and also to show combined green/blue prints before the laboratory moved to a Minimum Safe (MinSafe) Operations posture as a result of local Shelter In Place orders due to Covid-19.

42 ENGINEERING↗

DOE OSTI2020

Mixed Material Scintillator Systems (Quarterly Report FY20Q3)

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

DOE OSTI2020

FY2021 PAIN Portable Neutron Radiography Project October 1-15, 2020 Effort

At the start of FY2021, the PAIN portable neutron radiography project had approximately $26K of carryover funding from the FY2020 NA-22 DNN R&D PAIN budget that was used during the period October 1-15, 2020. During this period, the carryover funding was used to continue work on qualification of the Adelphi Technology DD108.2i neutron generator prior to returning for the DT upgrade, and further develop the neutron radiography modeling and simulation effort. This project is also being funded through NA-84 for FY2021.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

DOE OSTI2021

Advancing Fissile Materials Production Modeling with Adaptive Computing Environment and Simulations (ACES)

The Department of Energy’s National Nuclear Security Administration (DOE/NNSA) provides advanced capabilities to simulate the uranium enrichment process to support international negotiations on the peaceful use of nuclear energy. Uranium isotope separation centrifuges connected in a cascade configuration can produce the low-enriched uranium needed for nuclear power. However, those same centrifuges connected in a different configuration can also produce highly enriched uranium for nuclear weapons. Having the capability to assess cascade operations and identify nefarious activities promotes the peaceful uses of nuclear energy while restricting nuclear weapons proliferation. DNN R&D's Nonproliferation Stewardship Program Adaptive Computing Environment and Simulations (ACES) project is creating a modern, sustainable ecosystem of physics-based models and data-analytics tools that enables analysts to model uranium enrichment systems, simulate operational scenarios, and apply various policy options to explore potential outcomes.

07 ISOTOPE AND RADIATION SOURCES↗

DOE OSTI2020

Mixed Material Scintillator Systems Quarterly Report FY20Q4

This project aims to invent, model, and prototype architected multimaterial scintillator systems (AMSSs), a new class of radiation detectors that use heterogenous internal structures of different scintillating materials to detect additional properties of radiation. These internal structures can be produced using additive manufacture (AM, i.e. 3D printing) of scintillator, a currently emerging application of additive manufacture technology. AMSSs combine the low cost and complexity of conventional scintillation detectors with capabilities currently only available in more expensive and complex detectors. By identifying promising AMSS designs, this project will enable a new class of detectors to meet DNN’s mission needs for SNM detection. In past reporting, we have described this detector concept as “mixed material scintillator systems” (MMSS), but as part of preparing papers for publication we have concluded the term “architected multimaterial scintillator systems” (AMSS) better reflects the importance of structure in these detectors. We plan to use the AMSS acronym going forward. This quarter, we participated in an Independent Assessment of the work so far. The panelists on the review concluded that the AMSS concept has great potential and our work so far was effective at bringing that potential to light. They highly encouraged further work on this topic. The panelists also had several useful suggestions, which we took to heart. This review is described in more detail below. This quarter saw a resumption of work on the final task of this project, prototyping. In the first half of the quarter, this work focused on solving specific technical barriers to achieve 1x1x1 cm prints. In the second half, we focused on producing high-quality samples and making characterization measurements of those samples, with an eye on our deliverable report on the prototyping task. This report can be expected along with the end-of-year report on October 30. This project successfully spent the entirety of its remaining funds this quarter, except for $13k refunded to this project in the last days of the fiscal year due to LLNL cost adjustments. This project has captured the interest of many LLNL experts, and so we have been able to effectively use our budget to harness the available effort.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

DOE OSTI2020

Mixed Material Scintillator Systems (NA-22 Project End Report)

This project aimed to invent, model, and prototype architected multimaterial scintillator systems (AMSSs), a new class of radiation detectors that use heterogenous internal structures of different scintillating materials to detect additional properties of radiation. These internal structures can be produced using additive manufacture (AM, i.e. 3D printing) of scintillator, a currently emerging application of additive manufacture technology. AMSSs combine the low cost and complexity of conventional scintillation detectors with capabilities currently only available in more expensive and complex detectors. By identifying promising AMSS designs, this project enabled a new class of detectors to meet DNN’s mission needs for SNM detection.

36 MATERIALS SCIENCE↗

DOE OSTI2021

Development of a Neutron List Mode Collar (LMCL) and a List Mode Response Matrix Analysis Concept

This report was prepared for the Safeguards Program of the US Department of Energy’s (DOE’s) National Nuclear Security Administration (NNSA), Office of Defense Nuclear Nonproliferation Research and Development (DNN R&D). The report presents the development of the neutron nondestructive assay system, the List Mode Collar (LMCL) for the project OR16-List Mode for Collar-PD1La “List Mode Response Matrix for Advanced Correlated Neutron Analysis for Nuclear Safeguards.” The new list mode electronics developed under this project, and a spatial analysis concept called the List Mode Response Matrix are also described in this report. Analysis algorithms based on classification methods are published in a separate report. This research addresses the need to expand the capabilities of current nondestructive assay systems used for nuclear safeguards applications and considers the sustainability of safeguards technologies by the development of a “retrofit” concept using electronics based on modern standards. Furthermore, employing list mode data acquisition enabled the development of a spatial analysis concept and empirical measurement of a spatial response not previously used for safeguards neutron counting applications or measured in a traditional neutron collar detector and, therefore, provides new capability.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

DOE OSTI2021

Safeguards Technology for Thorium Fuel Cycles: Research and Development Needs Assessment and Recommendations

This report presents Safeguards Technology for Thorium Fuel Cycles: Research and Development Needs Assessment and Recommendations prepared for the National Nuclear Security Administration (NNSA) Office of Defense Nuclear Nonproliferation Research and Development (DNN R&D) Safeguards Program by a multilaboratory team from Oak Ridge National Laboratory, Los Alamos National Laboratory, and Y-12 National Security Complex. It documents key findings of a 2-year scoping study on “Safeguards Technology Needs Assessment for Leading Thorium Fuel Cycles” (project OR18-V-SG Tec Needs Th Fuel Cycles-PD1Lb).

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

DOE OSTI2021

Nu Tools: Exploring Practical Roles for Neutrinos in Nuclear Energy and Security

For decades, physicists have used neutrinos from nuclear reactors to advance basic science. These pursuits have inspired many ideas for application of neutrino detectors in nuclear energy and security. While developments in neutrino detectors are now making some of these ideas technically feasible, their value in the context of real needs and constraints has been unclear. This report seeks to help focus the picture of where neutrino technology may find practical roles in nuclear energy and security. This report is the final product of the Nu Tools study, commissioned in 2019 by the DOE National Nuclear Security Administration (NNSA) Office of Defense Nuclear Nonproliferation Research and Development (DNN R&D). The study was conducted over two years by a group of neutrino physicists and nuclear engineers. A central theme of the study and this report is that useful application of neutrinos will depend not only on advancing physics and technology but also on understanding the needs and constraints of potential end-users. The Study Approach emphasized broad end-user engagement. The major effort, undertaken from May to December 2020, was a series of engagements with the wider nuclear energy and security communities. Interviews with 41 experts revealed points of common understanding, which this report captures in three Cross-Cutting Findings, a Framework for Evaluating Utility, and seven Use Case Findings. The report concludes with two Recommendations. The findings and recommendations are summarized below. The respective ordering within each category does not represent a prioritization or implied value judgement.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗