Search NASA⌕ Search

SEARCH · Search NASA

Results for “deep”

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 OSTI2016

Neurogenesis Deep Learning: Extending deep networks to accommodate new classes

Neural machine learning methods, such as deep neural networks (DNN), have achieved remarkable success in a number of complex data processing tasks. These methods have arguably had their strongest impact on tasks such as image and audio processing – data processing domains in which humans have long held clear advantages over conventional algorithms. In contrast to biological neural systems, which are capable of learning continuously, deep artificial networks have a limited ability for incorporating new information in an already trained network. As a result, methods for continuous learning are potentially highly impactful in enabling the application of deep networks to dynamic data sets. Here, inspired by the process of adult neurogenesis in the hippocampus, we explore the potential for adding new neurons to deep layers of artificial neural networks in order to facilitate their acquisition of novel information while preserving previously trained data representations. Our results on the MNIST handwritten digit dataset and the NIST SD 19 dataset, which includes lower and upper case letters and digits, demonstrate that neurogenesis is well suited for addressing the stability-plasticity dilemma that has long challenged adaptive machine learning algorithms.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2019

Physics-Guided Deep Learning for Time-Series State Estimation Against False Data Injection Attacks

The modern power grid is a cyber-physical system. While the grid is becoming more intelligent with emerging sensing and communication techniques, new vulnerabilities are introduced and cyber security becomes a major concern. One type of cyber-attacks - False Data Injection Attacks (FDIAs) - exploits the limitations in traditional power system state estimation, and modifies system states without being detected. In this paper, we propose a physics-guided deep learning (PGDL) approach to defend against FDIAs. The PGDL takes real-time measurements as inputs to neural networks, outputs the estimated states, and reconstructs measurements considering power system physics. A deep recurrent neural network - Long Short-Term Memory (LSTM) - is employed to learn the temporal correlations among states. This hybrid learning model leads to a time-series state estimation method to defend against FDIAs. The simulation results using IEEE 14-bus test system demonstrate the accuracy and robustness of the proposed time-series state estimation under FDIAs.

Cyber security, state estimation, physics-guided d↗
DOE OSTI2017

Scaling Deep Learning on GPU and Knights Landing clusters

The speed of deep neural networks training has become a big bottleneck of deep learning research and development. For example, training GoogleNet by ImageNet dataset on one Nvidia K20 GPU needs 21 days. To speed up the training process, the current deep learning systems heavily rely on the hardware accelerators. However, these accelerators have limited on-chip memory compared with CPUs. To handle large datasets, they need to fetch data from either CPU memory or remote processors. We use both self-hosted Intel Knights Landing (KNL) clusters and multi-GPU clusters as our target platforms. From an algorithm aspect, current distributed machine learning systems are mainly designed for cloud systems. These methods are asynchronous because of the slow network and high fault-tolerance requirement on cloud systems. We focus on Elastic Averaging SGD (EASGD) to design algorithms for HPC clusters. Original EASGD used round-robin method for communication and updating. The communication is ordered by the machine rank ID, which is inefficient on HPC clusters. First, we redesign four efficient algorithms for HPC systems to improve EASGD's poor scaling on clusters. Async EASGD, Async MEASGD, and Hogwild EASGD are faster \textcolor{black}{than} their existing counterparts (Async SGD, Async MSGD, and Hogwild SGD, resp.) in all the comparisons. Finally, we design Sync EASGD, which ties for the best performance among all the methods while being deterministic. In addition to the algorithmic improvements, we use some system-algorithm codesign techniques to scale up the algorithms. By reducing the percentage of communication from 87% to 14%, our Sync EASGD achieves 5.3x speedup over original EASGD on the same platform. We get 91.5% weak scaling efficiency on 4253 KNL cores, which is higher than the state-of-the-art implementation.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2023

Degradation-related defect level in weathered silicon heterojunction modules characterized by deep level transient spectroscopy

Commercial silicon heterojunction photovoltaic modules, known as amorphous-silicon-based heterojunction with intrinsic thin-film layer (HIT) modules, show average degradation after 10 years in the field. HIT modules weathered outdoors in Colorado and Florida display mostly uniform decreases in intensity when mapped with photoluminescence (PL) imaging compared to a control module. Flash-table-based current-voltage curves show that degradation is dominated by voltage loss. Samples are cored from each of the modules, and deep level transient spectroscopy (DLTS) detects three electron-trap defect states in all modules with activation energies of electron emission from the defects of 0.07, 0.16, and 0.50 eV. DLTS measurements on the weathered modules show an additional deep-level, electron-trap defect state with an activation energy of 0.51 eV and a trap density of approximately 10 12 cm –3 . The capture rate is measured using varying short filling pulse times, and the resulting capture cross section is estimated to be 1.1x10 –16 cm 2 . The development of the weathering-related defect level correlates to decreases in carrier lifetime, PL intensity, and module voltage. Various depths of the space charge region are probed with increments in applied reverse bias and filling-pulse bias. Furthermore, this DLTS depth profiling shows a trend of trap density increasing with less applied reverse bias, suggesting that the weathering-related defect increases carrier recombination toward the interface between the bulk silicon wafer and the junction-forming amorphous-silicon passivation layers.

14 SOLAR ENERGY↗
DOE OSTI2018

Effects of Scavenging, Entrainment, and Aqueous Chemistry on Peroxides and Formaldehyde in Deep Convective Outflow Over the Central and Southeast United States

Deep convective transport of gaseous precursors to ozone (O3) and aerosols to the upper troposphere is affected by liquid phase and mixed-phase scavenging, entrainment of free tropospheric air and aqueous chemistry. The contributions of these processes are examined using aircraft measurements obtained in storm inflow and outflow during the 2012 Deep Convective Clouds and Chemistry (DC3) experiment combined with high-resolution (dx = 3 km) WRF-Chem simulations of a severe storm, an air mass storm, and a mesoscale convective system (MCS). The simulation results for the MCS suggest that formaldehyde (CH2O) is not retained in ice when cloud water freezes, in agreement with previous studies of the severe storm. By analyzing WRF-Chem trajectories, the effects of scavenging, entrainment, and aqueous chemistry on outflow mixing ratios of CH2O, methyl hydroperoxide (CH3OOH), and hydrogen peroxide (H2O2) are quantified. Liquid phase microphysical scavenging was the dominant process reducing CH2O and H2O2 outflow mixing ratios in all three storms. Aqueous chemistry did not significantly affect outflow mixing ratios of all three species. In the severe storm and MCS, the higher than expected reductions in CH3OOH mixing ratios in the storm cores were primarily due to entrainment of low-background CH3OOH. In the air mass storm, lower CH3OOH and H2O2 scavenging efficiencies (SEs) than in the MCS were partly due to entrainment of higher background CH3OOH and H2O2. Overestimated rain and hail production in WRF-Chem reduces the confidence in ice retention fraction values determined for the peroxides and CH2O.

Bela, Megan↗
DOE OSTI2017

A PDF-Based Parameterization of Subgrid-Scale Hydrometeor Transport in Deep Convection

Abstract A parameterization scheme is proposed for the subgrid-scale transport of hydrometeors in an assumed probability density function (PDF) scheme. Joint distributions of vertical velocity and hydrometeor mixing ratios are typically unknown, but marginal (1D) PDFs of these variables are available. The parameterization is developed using high-resolution simulations of continental and tropical deep convection. A 3D cloud-resolving model (CRM) providing benchmark solutions has a horizontal grid spacing of 250 m and employs the Morrison microphysics scheme, which treats prognostically mass and number mixing ratios for four types of precipitating hydrometeors (rain, graupel, snow, and ice) as well as cloud droplet number mixing ratio. The subgrid-scale hydrometeor transport scheme assumes input given in the form of marginal PDFs of vertical velocity and hydrometeor mixing ratios; in this study, these marginal distributions are provided by the cloud-resolving model. Conditional sampling and scaling are then applied to the marginal distributions to account for subplume correlations. The parameterized fluxes tested for four episodes of deep convection show good agreement with benchmark fluxes computed directly from the CRM output. The results demonstrate the potential use of the subgrid-scale hydrometeor transport scheme in an assumed PDF scheme to parameterize the covariances of vertical velocity and hydrometeor mixing ratios.

Wong, May↗
DOE OSTI2026

Data for FUN-PROSE: A Deep Learning Approach to Predict Condition-Specific Gene Expression in Fungi

mRNA levels of all genes in a genome is a critical piece of information defining the overall state of the cell in a given environmental condition. Being able to reconstruct such condition-specific expression in fungal genomes is particularly important to metabolically engineer these organisms to produce desired chemicals in industrially scalable conditions. Most previous deep learning approaches focused on predicting the average expression levels of a gene based on its promoter sequence, ignoring its variation across different conditions. Here we present FUN-PROSE—a deep learning model trained to predict differential expression of individual genes across various conditions using their promoter sequences and expression levels of all transcription factors. We train and test our model on three fungal species and get the correlation between predicted and observed condition-specific gene expression as high as 0.85. We then interpret our model to extract promoter sequence motifs responsible for variable expression of individual genes. We also carried out input feature importance analysis to connect individual transcription factors to their gene targets. A sizeable fraction of both sequence motifs and TF-gene interactions learned by our model agree with previously known biological information, while the rest corresponds to either novel biological facts or indirect correlations.

Genomics↗
DOE OSTI2022

Deep Realistic Extragalactic Model (DREaM) Galaxy Catalogs: Predictions for a Roman Ultra-deep Field

In the next decade, deep galaxy surveys from telescopes such as the James Webb Space Telescope and Roman Space Telescope will provide transformational data sets that will greatly enhance the understanding of galaxy formation during the epoch of reionization (EoR). In this work, we present the Deep Realistic Extragalactic Model (DREaM) for creating synthetic galaxy catalogs. Our model combines dark matter simulations, subhalo abundance matching and empirical models, and includes galaxy positions, morphologies, and spectral energy distributions. The resulting synthetic catalog extends to redshifts z ~ 12, and galaxy masses log 10 (M/M ⊙ ) = 5 covering an area of 1 deg 2 on the sky. We use DREaM to explore the science returns of a 1 deg 2 Roman ultra-deep field (UDF), and to provide a resource for optimizing ultra-deep survey designs. We find that a Roman UDF to ~30 m AB will potentially detect more than 10 6 M UV < -17 galaxies, with more than 10 4 at redshifts z > 7, offering an unparalleled data set for constraining galaxy properties during the EoR. Our synthetic catalogs and simulated images are made publicly available to provide the community with a tool to prepare for upcoming data.

79 ASTRONOMY AND ASTROPHYSICS↗
NASA NTRS2021

Scheduling the NASA Deep Space Network with Deep Reinforcement Learning

With three complexes spread evenly across the Earth, NASA’s Deep Space Network (DSN) is the primary means of communications as well as a significant scientific instrument for dozens of active missions around the world. A rapidly rising number of spacecraft and increasingly complex scientific instruments with higher bandwidth requirements have resulted in demand that exceeds the network’s capacity across its 12 antennae. The existing DSN scheduling process operates on a rolling weekly basis and is time-consuming; for a given week, generation of the final baseline schedule of spacecraft tracking passes takes roughly 5 months from the initial requirements submission deadline, with several weeks of peer-to-peer negotiations in between. This paper proposes a deep reinforcement learning (RL) approach to generate candidate DSN schedules from mission requests and spacecraft ephemeris data with demonstrated capability to address real-world operational constraints. A deep RL agent is developed that takes mission requests for a given week as input, and interacts with a DSN scheduling environment to allocate tracks such that its reward signal is maximized. A comparison is made between an agent trained using Proximal Policy Optimization and its random, untrained counterpart. The results represent a proof-of-concept that, given a well-shaped reward signal, a deep RL agent can learn the complex heuristics used by experts to schedule the DSN. A trained agent can potentially be used to generate candidate schedules to bootstrap the scheduling process and thus reduce the turnaround cycle for DSN scheduling.

Wilson, Brian↗
DOE OSTI2018

Automated correlative segmentation of large Transmission X-ray Microscopy (TXM) tomograms using deep learning

Highlights: • A unique automated segmentation approach for large 3D nanotomography datasets obtained by Transmission X-ray Microscopy (TXM) in an Al-Cu alloy. • Automated segmentation using a Convolutional Neural Network (CNN) architecture based on a deep learning approach was employed. • This extremely versatile technique is capable of emulating the manual segmentation process effectively. • Quantitative comparison between manual segmentation and CNN architecture established the accuracy of this automated technique. • Its ability to robustly process ultra-large volumes of data in relatively small time frames can exponentially accelerate tomographic data analysis. - Abstract: A unique correlative approach for automated segmentation of large 3D nanotomography datasets obtained using Transmission X-ray Microscopy (TXM) in an Al-Cu alloy has been introduced. Automated segmentation using a Convolutional Neural Network (CNN) architecture based on a deep learning approach was employed. This extremely versatile technique is capable of emulating the manual segmentation process effectively. Coupling this technique with post-scanning SEM imaging ensured precise estimation of 3D morphological parameters from nanotomography. The segmentation process as well as subsequent analysis was expedited by several orders of magnitude. Quantitative comparison between segmentation performed manually and using the CNN architecture established the accuracy of this automated technique. Its ability to robustly process ultra-large volumes of data in relatively small time frames can exponentially accelerate tomographic data analysis, possibly opening up novel avenues for performing 4D characterization experiments with finer time steps.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Bayesian Optimized Deep Ensemble for Uncertainty Quantification of Deep Neural Networks: a System Safety Case Study on Sodium Fast Reactor Thermal Stratification Modeling

Deep neural networks (DNNs) are increasingly important to scientific computing and engineering system simulations. Accurate uncertainty quantification (UQ) for DNNs is critical in safety-sensitive engineering domains. Traditional Deep Ensemble (DE) methods, while easy to implement, frequently suffer from poorly calibrated uncertainty estimates and limited predictive accuracy due to reliance on fixed architectures with varied weight initializations. To address these issues, we introduce a workflow that combines Bayesian Optimization (BO) and DE. The workflow is modular, scalable, and integrates parallel BO initialized with Sobol sequences to individually optimize the hyperparameters of each ensemble member. This method enhances ensemble diversity, improves predictive accuracy, and provides reliable uncertainty estimates. We evaluate the proposed BODE approach in a sodium fast reactor thermal stratification modeling case study, where we used a densely connected convolutional neural network to predict turbulent viscosity during the reactor transient with consideration of data noise. We benchmark its performance against several optimization approaches, including baseline deep ensemble, evolutionary algorithm-optimized ensemble, ensemble formed via random search combined with greedy selection, and a BO ensemble using random initialization. Here, our results demonstrate superior performance of the developed BODE approach. In noise-free scenarios, BODE notably reduces incorrect aleatoric uncertainty and significantly enhances predictive accuracy. Under conditions of 5% and 10% Gaussian noise, BODE adaptively quantifies uncertainty proportional to data noise, achieving up to an 80% reduction in root mean square error compared to baseline methods and producing well-calibrated prediction intervals.

Bayesian optimization↗
DOE OSTI2018

Deep crustal heating by neutrinos from the surface of accreting neutron stars

Here, we present a new mechanism for deep crustal heating in accreting neutron stars. Charged pions (π + ) are produced in nuclear collisions on the neutron star surface during active accretion and upon decay they provide a flux of neutrinos into the neutron star crust. For massive and/or compact neutron stars, neutrinos deposit ≈ 1 – 2 MeV of heat per accreted nucleon into the inner crust. The strength of neutrino heating is comparable to the previously known sources of deep crustal heating, such as from pycnonuclear fusion reactions, and is relevant for studies of cooling neutron stars. We model the thermal evolution of a transient neutron star in a low-mass x-ray binary, and in the particular case of the neutron star MXB 1659-29 we show that additional deep crustal heating requires a higher thermal conductivity for the neutron star inner crust. A better knowledge of pion-production cross sections near threshold would improve the accuracy of our predictions.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2019

Deep Convective Organization, Moisture Vertical Structure, and Convective Transition Using Deep-Inflow Mixing

It is an open question whether an integrated measure of buoyancy can yield a strong relation to precipitation across tropical land and ocean, across the seasonal and diurnal cycles, and for varying degrees of convective organization. Building on previous work, entraining plume buoyancy calculations reveal that differences in convective onset as a function of column water vapor (CWV) over land and ocean, as well as seasonally and diurnally over land, are largely due to variability in the contribution of lower-tropospheric humidity to the total column moisture. Over land, the relationship between deep convection and lower-free-tropospheric moisture is robust across all seasons and times of day, whereas the relation to boundary layer moisture is robust for the daytime only. Using S-band radar, these transition statistics are examined separately for mesoscale and smaller-scale convection. The probability of observing mesoscale convective systems sharply increases as a function of lower-free-tropospheric humidity. The consistency of this with buoyancy-based parameterization is examined for several mixing formulations. Mixing corresponding to deep inflow of environmental air into a plume that grows with height, which incorporates nearly equal weighting of boundary layer and free-tropospheric air, yields buoyancies consistent with the observed onset of deep convection across the seasonal and diurnal cycles in the Amazon. Furthermore, it provides relationships that are as strong or stronger for mesoscale-organized convection as for smaller-scale convection.

54 ENVIRONMENTAL SCIENCES↗
NASA NTRS1981

Radio frequency interference protection of communications between the Deep Space Network and deep space flight projects

The increasing density of electrical and electronic circuits in Deep Space Station systems for computation, control, and numerous related functions has combined with the extension of system performance requirements calling for higher speed circuitry along with broader bandwidths. This has progressively increased the number of potential sources of radio frequency interference inside the stations. Also, the extension of spectrum usage both in power and frequency as well as the greater density of usage at all frequencies for national and international satellite communications, space research, Earth resource operations and defense, and particularly the huge expansion of airborne electronic warfare and electronic countermeasures operations in the Mojave area have greatly increased the potential number and severity of radio frequency interference incidents. The various facets of this problem and the efforts to eliminate or minimize the impact of interference on Deep Space Network support of deep space flight projects are described.

Johnston, D. W. H.↗