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 487 records · Page 27

Stream Temperature Prediction in a Shifting Environment: Explaining the Influence of Deep Learning Architecture

Stream temperature is a fundamental control on ecosystem health. Recent efforts incorporating process guidance into deep learning models for predicting stream temperature have been shown to outperform existing statistical and physical models. This performance is in part because deep learning architectures can actively learn spatiotemporal relationships that govern how water and energy propagate through a river network. However, exploration of how spatiotemporal awareness and process guidance influence a model's generalizability under shifting environmental conditions such as climate change is limited. Here, we use Explainable Artificial Intelligence (XAI) to interrogate how differing deep learning architectures affect a model's learned spatial and temporal dependencies, and how those learned dependencies affect a model's ability to maintain high accuracy when applied to unseen environmental conditions. Using the Delaware River Basin in the northeastern United States as a test case, we compare two spatiotemporally aware process–guided deep learning models for predicting stream temperature (a recurrent graph convolution network—RGCN, and a temporal convolution graph model—Graph WaveNet). Both models achieve equally high predictive performance when testing data are well represented in the training data (test root mean squared errors of 1.64°C and 1.65°C); however, Graph WaveNet significantly outperforms RGCN in 4 out of 5 experiments where test partitions represent different types of unseen environmental conditions. XAI results show that the architecture of Graph WaveNet leads to learned spatial relationships with greater fidelity to physical processes, and that this fidelity improves the generalizability of the model when applied to shifting and/or unseen environmental conditions.

54 ENVIRONMENTAL SCIENCES↗

Deep levels in a-plane, high Mg-content Mg{sub x}Zn{sub 1-x}O epitaxial layers grown by molecular beam epitaxy

Deep level defects in n-type unintentionally doped a-plane Mg{sub x}Zn{sub 1-x}O, grown by molecular beam epitaxy on r-plane sapphire were fully characterized using deep level optical spectroscopy (DLOS) and related methods. Four compositions of Mg{sub x}Zn{sub 1-x}O were examined with x = 0.31, 0.44, 0.52, and 0.56 together with a control ZnO sample. DLOS measurements revealed the presence of five deep levels in each Mg-containing sample, having energy levels of E{sub c} - 1.4 eV, 2.1 eV, 2.6 V, and E{sub v} + 0.3 eV and 0.6 eV. For all Mg compositions, the activation energies of the first three states were constant with respect to the conduction band edge, whereas the latter two revealed constant activation energies with respect to the valence band edge. In contrast to the ternary materials, only three levels, at E{sub c} - 2.1 eV, E{sub v} + 0.3 eV, and 0.6 eV, were observed for the ZnO control sample in this systematically grown series of samples. Substantially higher concentrations of the deep levels at E{sub v} + 0.3 eV and E{sub c} - 2.1 eV were observed in ZnO compared to the Mg alloyed samples. Moreover, there is a general invariance of trap concentration of the E{sub v} + 0.3 eV and 0.6 eV levels on Mg content, while at least and order of magnitude dependency of the E{sub c} - 1.4 eV and E{sub c} - 2.6 eV levels in Mg alloyed samples.

36 MATERIALS SCIENCE↗

Characterization of the deep levels responsible for non-radiative recombination in InGaN/GaN light-emitting diodes

This paper presents an extensive investigation of the deep levels related to non-radiative recombination in InGaN/GaN light-emitting diodes (LEDs). The study is based on combined optical and deep-level transient spectroscopy measurements, carried out on LEDs with identical structure and with different values of the non-radiative recombination coefficient. Experimental data lead to the following, relevant, results: (i) LEDs with a high non-radiative recombination coefficient have a higher concentration of a trap (labeled as “e{sub 2}”) with an activation energy of 0.7 eV, which is supposed to be located close to/within the active region; (ii) measurements carried out with varying filling pulse duration suggest that this deep level behaves as a point-defect/dislocation complex. The Arrhenius plot of this deep level is critically compared with the previous literature reports, to identify its physical origin.

36 MATERIALS SCIENCE↗

Correlation of a generation-recombination center with a deep level trap in GaN

We report on the identification of a deep level trap centre which contributes to generation-recombination noise. A n-GaN epilayer, grown by MOCVD on sapphire, was measured by deep level transient spectroscopy (DLTS) and noise spectroscopy. DLTS found 3 well documented deep levels at E{sub c} − 0.26 eV, E{sub c} − 0.59 eV, and E{sub c} − 0.71 eV. The noise spectroscopy identified a generation recombination centre at E{sub c} − 0.65 ± 0.1 eV with a recombination lifetime of 65 μs at 300 K. This level is considered to be the same as the one at E{sub c} − 0.59 eV measured from DLTS, as they have similar trap densities and capture cross section. This result shows that some deep levels contribute to noise generation in GaN materials.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Assessment of deep levels with selenium concentration in Cd 1–x Zn x Te 1–y Se y room temperature detector materials

Incorporation of Se into Cd 1–x Zn x Te (CZT) to form the quaternary compound semiconductor Cd 1–x Zn x Te 1–y Se y (CZTS) has proven to be an effective solution for compensating the major flaws associated with CZT, including poor homogeneity and high concentrations of electronically active deep levels that limit the performance of CZT detectors. In order to investigate how deep levels are affected by the Se concentration in CZTS, we performed photoinduced current transient spectroscopy (PICTS) measurements on CZTS crystals grown by the traveling heater method (THM) with 10% atomic Zn and varying atomic percentage of Se from 1.5% to 7.0%. The PICTS scans for up to 4% Se showed an exponential reduction in the capture cross section of deep levels associated with Te secondary phases in conjunction with an increase in a deep level positioned near the mid-gap, which initially increases the electron trapping time before degrading again at higher Se concentrations. The PICTS peaks present in 7% Se were anomalous relative to the other crystals and are expected to originate from transition metal impurities found in the lower-purity CdSe precursor material.

36 MATERIALS SCIENCE↗

A deep learning approach to fast analysis of collective Thomson scattering spectra

Fast analysis of collective Thomson scattering ion acoustic wave features using a deep convolutional neural network model is presented. The network was trained from spectra to predict the plasma parameters, including ion velocities, population fractions, and ion and electron temperatures. A fully kinetic particle-in-cell simulation was used to model a laboratory astrophysics experiment and simulate a diagnostic image of the ion acoustic wave feature. Network predictions were compared with Bayesian inference of the plasma model parameters for both the simulated and experimentally measured images. Both approaches were fairly accurate predicting the simulated image and the network predictions matched a good portion of the Bayesian results for the experimentally measured image. The Bayesian approach is more robust to noise and motivates future work to train deep learning models with realistic noise. The advantage of the deep learning model is making thousands of predictions in a few hundred milliseconds, compared to a few seconds to minutes per prediction for the optimization and Bayesian approaches presented here. The results demonstrate promising capabilities of deep learning models to analyze Thomson data orders of magnitude faster than conventional methods when using the neural network for standalone analysis. If more rigorous analysis is needed, neural network predictions can be used to quickly initialize other optimization methods and increase chances of success. This is especially useful when the dataset becomes very large or highly dimensional and manually refining initial conditions for the entire dataset are no longer tractable.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Transport and chemistry of isoprene and its oxidation products in deep convective clouds

Deep convective clouds can transport trace gases from the planetary boundary layer into the upper troposphere where subsequent chemistry may impact aerosol particle formation and growth. In this modelling study, we investigate processes that affect isoprene and its oxidation products injected into the upper troposphere by an isolated deep convective cloud in the Amazon. We run a photochemical box model with coupled cloud microphysics along hundreds of individual air parcel trajectories sampled from a cloud-resolving model simulation of a convective event. The box model simulates gas-phase chemical reactions, gas scavenging by liquid and ice hydrometeors, and turbulent dilution inside a deep convective cloud. The results illustrate the potential importance of gas uptake to anvil ice in regulating the intensity of the isoprene oxidation and associated low volatility organic vapour concentrations in the outflow. Isoprene transport and fate also depends on the abundance of lightning-generated nitrogen oxide radicals (NO x = NO + NO 2 ). If gas uptake on ice is efficient and lightning activity is low, around 30% of the boundary layer isoprene will survive to the cloud outflow after approximately one hour of transport, while all the low volatile oxidation products will be scavenged by the cloud hydrometeors. If lightning NO x is abundant and gas uptake by ice is inefficient, then all isoprene will be oxidised during transport or in the immediate outflow region, while several low volatility isoprene oxidation products will have elevated concentrations in the cloud outflow. Reducing uncertainties associated with the uptake of vapours on ice hydrometeors, especially HO 2 and oxygenated organics, is essential to improve predictions of isoprene and its oxidation products in deep convective outflows and their potential contribution to new particle formation and growth.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Novel candidate taxa contribute to key metabolic processes in Fennoscandian Shield deep groundwaters

The continental deep biosphere contains a vast reservoir of microorganisms, although a large proportion of its diversity remains both uncultured and undescribed. In this study, the metabolic potential (metagenomes) and activity (metatranscriptomes) of the microbial communities in Fennoscandian Shield deep subsurface groundwaters were characterized with a focus on novel taxa. DNA sequencing generated 1270 de-replicated metagenome-assembled genomes and single-amplified genomes, containing 7 novel classes, 34 orders, and 72 families. The majority of novel taxa were affiliated with Patescibacteria, whereas among novel archaea taxa, Thermoproteota and Nanoarchaeota representatives dominated. Metatranscriptomes revealed that 30 of the 112 novel taxa at the class, order, and family levels were active in at least one investigated groundwater sample, implying that novel taxa represent a partially active but hitherto uncharacterized deep biosphere component. The novel taxa genomes coded for carbon fixation predominantly via the Wood–Ljungdahl pathway, nitrogen fixation, sulfur plus hydrogen oxidation, and fermentative pathways, including acetogenesis. These metabolic processes contributed significantly to the total community’s capacity, with up to 9.9% of fermentation, 6.4% of the Wood–Ljungdahl pathway, 6.8% of sulfur plus 8.6% of hydrogen oxidation, and energy conservation via nitrate (4.4%) and sulfate (6.0%) reduction. Key novel taxa included the UBA9089 phylum, with representatives having a prominent role in carbon fixation, nitrate and sulfate reduction, and organic and inorganic electron donor oxidation. These data provided insights into deep biosphere microbial diversity and their contribution to nutrient and energy cycling in this ecosystem.

Candidatus↗

Interactions among rooting traits for deep water and nitrogen uptake in upland and lowland ecotypes of switchgrass ( Panicum virgatum L.)

Abstract The response of plant growth and development to nutrient and water availability is an important adaptation for abiotic stress tolerance. Roots need to intercept both passing nutrients and water while foraging into new soil layers for further resources. Substantial amounts of nitrate can be lost in the field when leaching into groundwater, yet very little is known about how deep rooting affects this process. Here, we phenotyped root system traits and deep 15N nitrate capture across 1.5 m vertical profiles of solid media using tall mesocosms in switchgrass (Panicum virgatum L.), a promising cellulosic bioenergy feedstock. Root and shoot biomass traits, photosynthesis and respiration measures, and nutrient uptake and accumulation traits were quantified in response to a water and nitrate stress factorial experiment for switchgrass upland (VS16) and lowland (AP13) ecotypes. The two switchgrass ecotypes shared common plastic abiotic responses to nitrogen (N) and water availability, and yet had substantial genotypic variation for root and shoot traits. A significant interaction between N and water stress combination treatments for axial and lateral root traits represents a complex and shared root development strategy for stress mitigation. Deep root growth and 15N capture were found to be closely linked to aboveground growth. Together, these results represent the wide genetic pool of switchgrass and show that deep rooting promotes nitrate capture, plant productivity, and sustainability.

59 BASIC BIOLOGICAL SCIENCES↗

Deep Learning for Automated Extraction of Primary Sites from Cancer Pathology Reports

Pathology reports are a primary source of information for cancer registries which process high volumes of free-text reports annually. Information extraction and coding is a manual, labor-intensive process. Here in this study we investigated deep learning and a convolutional neural network (CNN), for extracting ICDO- 3 topographic codes from a corpus of breast and lung cancer pathology reports. We performed two experiments, using a CNN and a more conventional term frequency vector approach, to assess the effects of class prevalence and inter-class transfer learning. The experiments were based on a set of 942 pathology reports with human expert annotations as the gold standard. CNN performance was compared against a more conventional term frequency vector space approach. We observed that the deep learning models consistently outperformed the conventional approaches in the class prevalence experiment, resulting in micro and macro-F score increases of up to 0.132 and 0.226 respectively when class labels were well populated. Specifically, the best performing CNN achieved a micro-F score of 0.722 over 12 ICD-O-3 topography codes. Transfer learning provided a consistent but modest performance boost for the deep learning methods but trends were contingent on CNN method and cancer site. Finally, these encouraging results demonstrate the potential of deep learning for automated abstraction of pathology reports.

59 BASIC BIOLOGICAL SCIENCES↗

Accelerating Hydraulic Fracture Imaging by Deep Transfer Learning

Deep transfer learning has a great success story in computer vision (CV), natural language processing, and many other fields. In this communication, we are going to push forward the deep transfer learning to the hydraulic fracture imaging problem by proposing a two-step approach: 1) train a convolutional neural network (CNN) to reconstruct target geometries by a relatively large amount of approximated field patterns generated from a simplified model and 2) fine-tune the top layers of transferred CNN by a small amount of true field patterns generated through a full model. The advantages include the rapid generation of large amount of data through the simplified model and the high reconstruction accuracy through a careful design of the deep transfer learning. Here, the CNN trained through the deep transfer learning can accurately reconstruct the lateral extent and direction of fractures with unseen conductivity and white Gaussian noise, showing a notable acceleration/accuracy enhancement over the previous CNN trained by mixing a small number of true data into the approximated data as data augmentation.

02 PETROLEUM↗

Deep Reinforcement Learning based Model-free On-line Dynamic Multi-Microgrid Formation to Enhance Resilience

Multi-microgrid formation (MMGF) is a promising solution for enhancing power system resilience. This paper proposes a new deep reinforcement learning (RL) based model-free on-line dynamic MMGF scheme. Additionally, the dynamic MMGF problem is formulated as a Markov decision process, and a complete deep RL framework is specially designed for the topologytransformable micro-grids. In order to reduce the large action space caused by flexible switch operations, a topology transformation method is proposed and an action-decoupling Q-value is applied. Then, a convolutional neural network (CNN) based multi-buffer double deep Q-network (CM-DDQN) is developed to further improve the learning ability of the original DQN method. The proposed deep RL method provides real-time computing to support the on-line dynamic MMGF scheme, and the scheme handles a long-term resilience enhancement problem using an adaptive on-line MMGF to defend changeable conditions. The effectiveness of the proposed method is validated using a 7-bus system and the IEEE 123-bus system. The results show strong learning ability, timely response for varying system conditions and convincing resilience enhancement.

24 POWER TRANSMISSION AND DISTRIBUTION↗

How Much Chemistry Does a Deep Neural Network Need to Know to Make Accurate Predictions?

In the last few years, we have seen the rise of deep learning applications in a broad range of computational chemistry research problems. Using human-engineered chemical features, such as molecular descriptors and fingerprints, deep learning models have shown similar, if not better performance that most traditional machine learning algorithms. Recently, we reported on the development of Chemception, a deep convolutional neural network (CNN) architecture for general-purpose small molecule property prediction. On average, Chemception matched the performance of expert-developed QSAR/QSPR models trained on chemical features (molecular fingerprints), despite that it was trained on just 2D images of molecular drawings with minimal chemical information. Here, we investigate the effects of systematically removing and adding basic chemical information to the image channels of the 2D images used to train Chemception. By augmenting our images with only 3 additional basic chemical information, we demonstrate the improvement of Chemception performance – that it is now more accurate than contemporary deep learning models trained on ECFP fingerprints for the prediction of toxicity, activity and solvation free energy, as well as physics-based free energy simulation methods for computing solvation properties. By altering the chemical information content in the image channels, and examining the resulting performance of Chemception, we also identify to two different “learning patterns” in toxicity/activity as compared to solvation free energy, and it parallels the fundamental differences in contemporary chemistry research for predicting toxicity/activity and solvation free energy.

Artificial Intelligence, Cheminformatics, Computat↗

A Combined Computer Vision and Deep Learning Approach for Rapid Drone-Based Optical Characterization of Parabolic Troughs

Optical accuracy is a primary driver of parabolic trough concentrating solar power (CSP) plant performance, but can be damaged by wind loads, gravity, error during installation, and regular plant operation. Collecting and analyzing optical measurements over an entire operating parabolic trough plant is difficult, given the large scale of typical installations. Distant Observer, a software tool developed at the National Renewable Energy Laboratory, uses images of the absorber tube reflected in the collector mirror to measure both surface slope in the parabolic mirror and offset of the absorber tube from the ideal focal point. This technology has been adapted for fast data collection using low-cost commercial drones, but until recently still required substantial human labor to process large amounts of data. A new method leveraging advanced deep learning and computer vision tools can drastically reduce the time required to process images. This new method addresses the primary analysis bottleneck, identifying featureless, reflective mirror corner points to a high degree of accuracy. Recent work has shown promising results using computer vision methods. The combined deep learning and computer vision approach presented here proved highly effective and has the potential to further automate data collection and analysis, making the tool more robust. The method presented in this paper automatically identified 74.3% of mirror corners within 2 pixels of their manually marked counterparts and 91.9% within 3 pixels. This level of accuracy is sufficient for practical Distant Observer analysis within a target uncertainty. A commercial drone collected video of over 100 parabolic trough modules at an operating CSP plant to demonstrate the deep learning and computer vision method's usefulness in processing large amounts of data. These troughs were successfully analyzed using Distant Observer, paired with the new deep learning and computer vision algorithm, and can provide plant operators and trough designers with valuable insight about plant performance, operating strategies, and plant-wide optical error trends.

computer vision↗

Deep electron levels in undoped polycrystalline CdTe annealed in liquid Cd

The method of deep-level transient spectroscopy (DLTS) is used to determine the set of deep electron levels in undoped n-CdTe polycrystals grown by chemical synthesis from vapor phase in highly nonequilibrium conditions and then subjected to annealing in liquid Cd. After annealing, the electron concentration is found to increase from {approx}10{sup 8} to 10{sup 15} cm{sup -3}. Electron traps with deep levels E{sub 1} = 0.84 {+-} 0.03 eV and E{sub 2} = 0.71 {+-} 0.02 eV with total concentration {approx}10{sup 14} cm{sup -3} are dominant in the DLTS spectrum. The role of grain boundaries and intrinsic point defects in formation of deep-level centers and their effect on the value of conductivity after annealing are discussed. A relation between the level E{sub 1} and complexes of intrinsic point defects and relation of the level E{sub 2} to the Cd{sub i} point defect are considered.

36 MATERIALS SCIENCE↗

Effect of Dislocation-related Deep Levels in Heteroepitaxial InGaAs/GaAs and GaAsSb/GaAs p–i–n Structures on the Relaxation time of Nonequilibrium Carriers

The results of an experimental study of the capacitance–voltage (C–V) characteristics and deep-level transient spectroscopy (DLTS) spectra of p{sup +}–p{sup 0}–i–n{sup 0} homostructures based on undoped dislocationfree GaAs layers and InGaAs/GaAs and GaAsSb/GaAs heterostructures with homogeneous networks of misfit dislocations, all grown by liquid-phase epitaxy (LPE), are presented. Deep-level acceptor defects identified as HL2 and HL5 are found in the epitaxial p{sup 0} and n{sup 0} layers of the GaAs-based structure. The electron and hole dislocation-related deep levels, designated as, respectively, ED1 and HD3, are detected in InGaAs/GaAs and GaAsSb/GaAs heterostructures. The following hole trap parameters: thermal activation energies (E{sub t}), capture cross sections (σ{sub p}), and concentrations (N{sub t}) are calculated from the Arrhenius dependences to be E{sub t} = 845 meV, σ{sub p} = 1.33 × 10{sup –12} cm{sup 2}, N{sub t} = 3.80 × 10{sup 14} cm{sup –3} for InGaAs/GaAs and E{sub t} = 848 meV, σ{sub p} = 2.73 × 10{sup –12} cm{sup 2}, N{sub t} = 2.40 × 10{sup 14} cm{sup –3} for GaAsSb/GaAs heterostructures. The concentration relaxation times of nonequilibrium carriers are estimated for the case in which dislocation-related deep acceptor traps are involved in this process. These are 2 × 10{sup –10} s and 1.5 × 10{sup –10} s for, respectively, the InGaAs/GaAs and GaAsSb/GaAs heterostructures and 1.6 × 10{sup –6} s for the GaAs homostructures.

36 MATERIALS SCIENCE↗

Identifying Genomic Islands with Deep Neural Networks

Background Horizontal gene transfer is the main source of adaptability for bacteria, through which genes are obtained from different sources including bacteria, archaea, viruses, and eukaryotes. This process promotes the rapid spread of genetic information across lineages, typically in the form of clusters of genes referred to as genomic islands (GIs). Different types of GIs exist, and are often classified by the content of their cargo genes or their means of integration and mobility. While various computational methods have been devised to detect different types of GIs, no single method is capable of detecting all types. Results We propose a method, which we call Shutter Island, that uses a deep learning model (Inception V3, widely used in computer vision) to detect genomic islands. The intrinsic value of deep learning methods lies in their ability to generalize. Via a technique called transfer learning, the model is pre-trained on a large generic dataset and then re-trained on images that we generate to represent genomic fragments. We demonstrate that this image-based approach generalizes better than the existing tools. Conclusions We used a deep neural network and an image-based approach to detect the most out of the correct GI predictions made by other tools, in addition to making novel GI predictions. The fact that the deep neural network was re-trained on only a limited number of GI datasets and then successfully generalized indicates that this approach could be applied to other problems in the field where data is still lacking or hard to curate.

Computer Vision↗

Water mass age structures the auxiliary metabolic gene content of free-living and particle-attached deep ocean viral communities

Viruses play important roles in the ocean’s biogeochemical cycles. Yet, deep ocean viruses are one of the most under-explored fractions of the global biosphere. Little is known about the environmental factors that control the composition and functioning of their communities or how they interact with their free-living or particle-attached microbial hosts. We analysed 58 viral communities associated with size-fractionated free-living (0.2–0.8 μm) and particle-attached (0.8–20 μm) cellular metagenomes from bathypelagic (2150–4018 m deep) microbiomes obtained during the Malaspina expedition. These metagenomes yielded 6631 viral sequences, 91% of which were novel, and 67 represented high-quality genomes. Taxonomic classification assigned 53% of the viral sequences to families of tailed viruses from the order Caudovirales. Computational host prediction associated 886 viral sequences to dominant members of the deep ocean microbiome, such as Alphaproteobacteria (284), Gammaproteobacteria (241), SAR324 (23), Marinisomatota (39), and Chloroflexota (61). Free-living and particle-attached viral communities had markedly distinct taxonomic composition, host prevalence, and auxiliary metabolic gene content, which led to the discovery of novel viral-encoded metabolic genes involved in the folate and nucleotide metabolisms. Water mass age emerged as an important factor driving viral community composition. We postulated this was due to changes in quality and concentration of dissolved organic matter acting on the host communities, leading to an increase of viral auxiliary metabolic genes associated with energy metabolism among older water masses. These results shed light on the mechanisms by which environmental gradients of deep ocean ecosystems structure the composition and functioning of free-living and particle-attached viral communities.

59 BASIC BIOLOGICAL SCIENCES↗