Search NASA⌕ Search

SEARCH · Search NASA

Results for “Deep ensemble”

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 253 records · Page 14
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Accurate and Rapid Forecasts for Geologic Carbon Storage via Learning-Based Inversion-Free Prediction

Carbon capture and storage (CCS) is one approach being studied by the U.S. Department of Energy to help mitigate global warming. The process involves capturing CO 2 emissions from industrial sources and permanently storing them in deep geologic formations (storage reservoirs). However, CCS projects generally target “green field sites,” where there is often little characterization data and therefore large uncertainty about the petrophysical properties and other geologic attributes of the storage reservoir. Consequently, ensemble-based approaches are often used to forecast multiple realizations prior to CO 2 injection to visualize a range of potential outcomes. In addition, monitoring data during injection operations are used to update the pre-injection forecasts and thereby improve agreement between forecasted and observed behavior. Thus, a system for generating accurate, timely forecasts of pressure buildup and CO 2 movement and distribution within the storage reservoir and for updating those forecasts via monitoring measurements becomes crucial. This study proposes a learning-based prediction method that can accurately and rapidly forecast spatial distribution of CO 2 concentration and pressure with uncertainty quantification without relying on traditional inverse modeling. The machine learning techniques include dimension reduction, multivariate data analysis, and Bayesian learning. The outcome is expected to provide CO 2 storage site operators with an effective tool for timely and informative decision making based on limited simulation and monitoring data.

58 GEOSCIENCES↗
NASA NTRS2019

Exploring the Utility of Machine Learning-Based Passive Microwave Brightness Temperature Data Assimilation over Terrestrial Snow in High Mountain Asia

This study explores the use of a support vector machine (SVM) as the observation operator within a passive microwave brightness temperature data assimilation framework (herein SVM-DA) to enhance the characterization of snow water equivalent (SWE) over High Mountain Asia (HMA). A series of synthetic twin experiments were conducted with the NASA Land Information System (LIS) at a number of locations across HMA. Overall, the SVM-DA framework is effective at improving SWE estimates (~70% reduction in RMSE relative to the Open Loop) for SWE depths less than 200 mm during dry snowpack conditions. The SVM-DA framework also improves SWE estimates in deep, wet snow (~45% reduction in RMSE) when snow liquid water is well estimated by the land surface model, but can lead to model degradation when snow liquid water estimates diverge from values used during SVM training. In particular, two key challenges of using the SVM-DA framework were observed over deep, wet snowpacks. First, variations in snow liquid water content dominate the brightness temperature spectral difference (TB) signal associated with emission from a wet snowpack, which can lead to abrupt changes in SWE during the analysis update. Second, the ensemble of SVM-based predictions can collapse (i.e., yield a near-zero standard deviation across the ensemble) when prior estimates of snow are outside the range of snow inputs used during the SVM training procedure. Such a scenario can lead to the presence of spurious error correlations between SWE and TB, and as a consequence, can result in degraded SWE estimates from the analysis update. These degraded analysis updates can be largely mitigated by applying rule-based approaches. For example, restricting the SWE update when the standard deviation of the predicted TB is greater than 0.05 K helps prevent the occurrence of filter divergence. Similarly, adding a thin layer (i.e., 5 mm) of SWE when the synthetic TB is larger than 5 K can improve SVM-DA performance in the presence of a precipitation dry bias. The study demonstrates that a carefully constructed SVM-DA framework cognizant of the inherent limitations of passive microwave-based SWE estimation holds promise for snow mass data assimilation.

Kwon, Yonghwan↗
DOE OSTI2025

Feature learning and generalization in deep networks with orthogonal weights

Fully-connected deep neural networks with weights initialized from independent Gaussian distributions can be tuned to criticality, which prevents the exponential growth or decay of signals propagating through the network. However, such networks still exhibit fluctuations that grow linearly with the depth of the network, which may impair the training of networks with width comparable to depth. We show analytically that rectangular networks with tanh activations and weights initialized from the ensemble of orthogonal matrices have corresponding preactivation fluctuations which are independent of depth, to leading order in inverse width. Moreover, we demonstrate numerically that, at initialization, all correlators involving the neural tangent kernel (NTK) and its descendants at leading order in inverse width—which govern the evolution of observables during training—saturate at a depth of ~20, rather than growing without bound as in the case of Gaussian initializations. We speculate that this structure preserves finite-width feature learning while reducing overall noise, thus improving both generalization and training speed in deep networks with depth comparable to width. We provide some experimental justification by relating empirical measurements of the NTK to the superior performance of deep non-linear orthogonal networks trained under full-batch gradient descent on the MNIST and CIFAR-10 classification tasks.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Advancing Organized Convection Representation in the Unified Model: Implementing and Enhancing Multiscale Coherent Structure Parameterization

To address the effect of stratiform latent heating on meso- to large-scale circulations, an enhanced implementation of the Multiscale Coherent Structure Parameterization (MCSP) is developed for the Met Office Unified Model. MCSP represents the top-heavy stratiform latent heating from under-resolved organized convection in general circulation models. We couple the MCSP with a mass-flux convection scheme (CoMorph-A) to improve storm lifecycle continuity. The improved MCSP trigger is specifically designed for mixed-phase deep convective cloud, combined with a background vertical wind shear, both known to be crucial for stratiform development. We also test a cloud top temperature dependent convective-stratiform heating partitioning, in contrast to the earlier fixed partitioning. Assessments from ensemble weather forecasts and decadal simulations demonstrate that MCSP directly reduces cloud deepening and precipitation areas by moderating mesoscale circulations. Indirectly, it amends tropical precipitation biases, notably correcting dry and wet biases over India and the Indian Ocean, respectively. Remarkably, the scheme outperforms a climate model ensemble by improving seasonal precipitation cycle predictions in these regions. The scheme also improves Madden-Julian Oscillation (MJO) spectra, achieving better alignment with observational and reanalysis data by intensifying the simulated MJO over the Indian Ocean during phases 4 to 5. However, the scheme increases precipitation overestimation over the Western Pacific. Shifting from fixed to temperature-dependent convective-stratiform partitioning reduces the Pacific precipitation overestimation and further improves the seasonal cycle in India. Spatially correlated biases highlight the necessity for advances beyond deterministic approaches to align MCSP with environmental conditions.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2025

Precision calibration of calorimeter signals in the ATLAS experiment using an uncertainty-aware neural network

The ATLAS experiment at the Large Hadron Collider explores the use of modern neural networks for a multi-dimensional calibration of its calorimeter signal defined by clusters of topologically connected cells (topo-clusters). The Bayesian neural network (BNN) approach not only yields a continuous and smooth calibration function that improves performance relative to the standard calibration but also provides uncertainties on the calibrated energies for each topo-cluster. The results obtained by using a trained BNN are compared to the standard local hadronic calibration and to a calibration provided by training a deep neural network. The uncertainties predicted by the BNN are interpreted in the context of a fractional contribution to the systematic uncertainties of the trained calibration. They are also compared to uncertainty predictions obtained from an alternative estimator employing repulsive ensembles.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2024

Deep-Learning-derived Boundary Layer Height from Meteorological Data over the SGP, GOAMAZON, CACTI

The planetary boundary-layer (PBL) height (PBLH) is an important parameter for various meteorological and climate studies. This study presents a multi-structure deep neural network (DNN) model, designed to estimate PBLH by integrating morning temperature profiles with surface meteorological observations. The DNN model is developed by leveraging a rich data set of PBLH derived from long-standing radiosonde records and augmented with high-resolution micropulse lidar and Doppler lidar observations. We access the performance of the DNN with an ensemble of 10 members, each featuring distinct hidden layer structures, which collectively yield a robust 27-year PBLH data set over the Southern Great Plains from 1994 to 2020. The influence of various meteorological factors on PBLH is rigorously analyzed through the importance test. Moreover, the DNN model's accuracy is evaluated against radiosonde observations and juxtaposed with conventional remote-sensing methodologies, including Doppler lidar, ceilometer, Raman lidar, and micropulse lidar. The DNN model exhibits reliable performance across diverse conditions and demonstrates lower biases relative to remote-sensing methods. In addition, the DNN model, originally trained over a plain region, demonstrates remarkable adaptability when applied to the heterogeneous terrains and climates encountered during the GoAmazon (tropical rainforest) and CACTI (middle-latitude mountain) campaigns. These findings demonstrate the effectiveness of deep learning models in estimating PBLH, enhancing our understanding of boundary-layer dynamics with implications for enhancing the representation of PBL in weather forecasting and climate modeling.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Report on the AAPM grand challenge on deep generative modeling for learning medical image statistics

Abstract Background The findings of the 2023 AAPM Grand Challenge on Deep Generative Modeling for Learning Medical Image Statistics are reported in this Special Report. Purpose The goal of this challenge was to promote the development of deep generative models for medical imaging and to emphasize the need for their domain‐relevant assessments via the analysis of relevant image statistics. Methods As part of this Grand Challenge, a common training dataset and an evaluation procedure was developed for benchmarking deep generative models for medical image synthesis. To create the training dataset, an established 3D virtual breast phantom was adapted. The resulting dataset comprised about 108 000 images of size 512 512. For the evaluation of submissions to the Challenge, an ensemble of 10 000 DGM‐generated images from each submission was employed. The evaluation procedure consisted of two stages. In the first stage, a preliminary check for memorization and image quality (via the Fréchet Inception Distance [FID]) was performed. Submissions that passed the first stage were then evaluated for the reproducibility of image statistics corresponding to several feature families including texture, morphology, image moments, fractal statistics, and skeleton statistics. A summary measure in this feature space was employed to rank the submissions. Additional analyses of submissions was performed to assess DGM performance specific to individual feature families, the four classes in the training data, and also to identify various artifacts. Results Fifty‐eight submissions from 12 unique users were received for this Challenge. Out of these 12 submissions, 9 submissions passed the first stage of evaluation and were eligible for ranking. The top‐ranked submission employed a conditional latent diffusion model, whereas the joint runners‐up employed a generative adversarial network, followed by another network for image superresolution. In general, we observed that the overall ranking of the top 9 submissions according to our evaluation method (i) did not match the FID‐based ranking, and (ii) differed with respect to individual feature families. Another important finding from our additional analyses was that different DGMs demonstrated similar kinds of artifacts. Conclusions This Grand Challenge highlighted the need for domain‐specific evaluation to further DGM design as well as deployment. It also demonstrated that the specification of a DGM may differ depending on its intended use.

Radiology, Nuclear Medicine & Medical Imaging↗
DOE OSTI2025

Spatially resolved charge-transfer kinetics at the quantum dot–microbe interface using fluorescence lifetime imaging microscopy

Integrating the optoelectronic properties of quantum dots (QDs) with biological enzymatic systems to form microbe-semiconductor biohybrids offers promising prospects for both solar-to-chemical conversion and light-modulated biochemical processes. Developing these nano–bio hybrid systems necessitates a deep understanding of charge-transfer dynamics at the nano–bio interface. Photoexcited carrier transfer from QDs to microbes is driven by complex interactions, with emerging insights into the relevant thermodynamic and kinetic factors. The heterogeneities of both microbes and QD ensembles pose significant challenges in mechanistic understanding, which is critical for designing advanced nano–bio hybrids. We used fluorescence lifetime imaging microscopy to analyze charge transfer between a CdSe QD film andShewanella oneidensismicrobes. We correlated the spatiotemporal fluorescence data with an analytical model. Our analysis revealed two distinct distributions of QD de-excitation pathways. The characteristics of these distributions: 1) a faster transfer rate ( k ¯ E T 1 = 1.5 10 9 s - 1 ), with a lower acceptor number ( N ¯ a 1 = 0.03 ) and 2) a slower transfer rate ( k ¯ E T 2 = 4.1 10 8 s - 1 ) with a higher acceptor number ( N ¯ a 2 = 0.18 ). We assign these distributions to the indirect and direct electron transfer mechanisms, respectively. Our findings demonstrate how spectroscopic imaging can uncover fundamental electron transfer mechanisms at complex interfaces, offering valuable design principles for future nano–bio hybrids.

Science & Technology - Other Topics↗
DOE OSTI2026

Stochastic mean-field theory and applications to multinucleon transfer and kinetic energy dissipation processes in heavy-ion collisions

In this Review article, a brief description of the stochastic mean-field (SMF) theory for describing reaction dynamics in low-energy heavy-ion collisions at bombarding energies in the vicinity of the Coulomb barrier is presented. In these collisions, as a result of strong Pauli blocking, binary nucleon collisions do not have a significant effect on the dissipation and fluctuations. At low energies, the mean-field fluctuations, due to initial correlations, have a dominant effect on fluctuations of macroscopic variables. The SMF theory proposes the determination of an ensemble of single-particle density matrices by specifying random initial fluctuations according to a distribution law. Employing an ensemble of single-particle density matrices, not only the mean values but also the distribution functions of the one-body observables can be determined. If the di-nuclear structure is maintained in heavy-ion collisions, such as deep inelastic collisions and fast quasi-fission reactions, a much simpler description of the reaction mechanism can be derived in terms of several macroscopic variables such as mass and charge asymmetry, and relative linear and relative angular momentum. In this case, by geometric projection of the SMF equations, it is possible to derive the quantal Langevin equations for macroscopic variables. As an application of quantal transport description, an analysis of multinucleon transfers and kinetic energy dissipation and fluctuations is presented for selected quasi-fission reactions.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2024

Deep-learning-derived planetary boundary layer height from conventional meteorological measurements

Abstract. The planetary boundary layer (PBL) height (PBLH) is an important parameter for various meteorological and climate studies. This study presents a multi-structure deep neural network (DNN) model, which can estimate PBLH by integrating the morning temperature profiles and surface meteorological observations. The DNN model is developed by leveraging a rich dataset of PBLH derived from long-standing radiosonde records augmented with high-resolution micro-pulse lidar and Doppler lidar observations. We access the performance of the DNN with an ensemble of 10 members, each featuring distinct hidden-layer structures, which collectively yield a robust 27-year PBLH dataset over the southern Great Plains from 1994 to 2020. The influence of various meteorological factors on PBLH is rigorously analyzed through the importance test. Moreover, the DNN model's accuracy is evaluated against radiosonde observations and juxtaposed with conventional remote sensing methodologies, including Doppler lidar, ceilometer, Raman lidar, and micro-pulse lidar. The DNN model exhibits reliable performance across diverse conditions and demonstrates lower biases relative to remote sensing methods. In addition, the DNN model, originally trained over a plain region, demonstrates remarkable adaptability when applied to the heterogeneous terrains and climates encountered during the GoAmazon (Green Ocean Amazon; tropical rainforest) and CACTI (Cloud, Aerosol, and Complex Terrain Interactions; middle-latitude mountain) campaigns. These findings demonstrate the effectiveness of deep learning models in estimating PBLH, enhancing our understanding of boundary layer processes with implications for improving the representation of PBL in weather forecasting and climate modeling.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2022

Technology Pathways Could Help Drive the U.S. West Coast Grid's Exposure to Hydrometeorological Uncertainty

Abstract Previous studies investigating deep decarbonization of bulk electric power systems and wholesale electricity markets have not sufficiently explored how future grid pathways could affect the grid's vulnerability to hydrometeorological uncertainty on multiple timescales. Here, we employ a grid operations model and a large synthetic weather ensemble to “stress test” a range of future grid pathways for the U.S. West Coast developed by ReEDS, a well‐known capacity planning model. Our results show that gradual changes in the underlying capacity mix from 2020 to 2050 can cause significant “re‐ranking” of weather years in terms of annual wholesale electricity prices (with “good” years becoming bad, and vice versa). Nonetheless, we find the highest and lowest ranking price years in terms of average electricity price remain mostly tied to extremes in hydropower availability (streamflow) and load (summer temperatures), with the strongest sensitivities related to drought. Seasonal dynamics seen today involving spring snowmelt and hot, dry summers remain well‐defined out to 2050. In California, future supply shortfalls in our model are concentrated in the evening and occur mostly during periods of high temperature anomalies in late summer months and in late winter; in the Pacific Northwest, supply shortfalls are much more strongly tied to negative streamflow anomalies. Under our more robust sampling of stationary hydrometeorological uncertainty, we also find that the ratio of dis‐patchable thermal (i.e., natural gas) capacity to wind and solar required to ensure grid reliability can differ significantly from values reported by ReEDS.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗
DOE OSTI2022

Impact of heavy-quark production measurements in the CT18 global QCD analysis of PDFs

We discuss the impact of heavy-flavor production measurements in semi-inclusive deep inelastic scattering at HERA on the CTEQ-TEA PDFs. In particular, we study the impact of the latest charm and bottom production measurements from the H1 and ZEUS collaborations on the gluon, and the interplay of these measurements with the data ensemble of the recent CT18 global QCD analysis.

Guzzi, Marco↗
NASA NTRS2008

Evaluation of Long-Term Cloud-Resolving Model Simulations Using Satellite Radiance Observations and Multi-Frequency Satellite Simulators

This paper proposes a methodology known as the Tropical Rainfall Measuring Mission (TRMM) Triple-Sensor Three-step Evaluation Framework (T3EF) for the systematic evaluation of precipitating cloud types and microphysics in a cloud-resolving model (CRM). T3EF utilizes multi-frequency satellite simulators and novel statistics of multi-frequency radiance and backscattering signals observed from the TRMM satellite. Specifically, T3EF compares CRM and satellite observations in the form of combined probability distributions of precipitation radar (PR) reflectivity, polarization-corrected microwave brightness temperature (Tb), and infrared Tb to evaluate the candidate CRM. T3EF is used to evaluate the Goddard Cumulus Ensemble (GCE) model for cases involving the South China Sea Monsoon Experiment (SCSMEX) and Kwajalein Experiment (KWAJEX). This evaluation reveals that the GCE properly captures the satellite-measured frequencies of different precipitating cloud types in the SCSMEX case but underestimates the frequencies of deep convective and deep stratiform types in the KWAJEX case. Moreover, the GCE tends to simulate excessively large and abundant frozen condensates in deep convective clouds as inferred from the overestimated GCE-simulated radar reflectivities and microwave Tb depressions. Unveiling the detailed errors in the GCE s performance provides the best direction for model improvements.

Matsui, Toshihisa↗
NASA NTRS1985

Response of cumulus clouds to large-scale forcing and cumulus parameterization

Using the cumulus ensemble model (Soong and Ogura, 1980), statistical properties of cumulus clouds that occur in response to the imposed large scale forcing were investigated both in the tropical and midlatitude situtations. A strong drying process was found to occur in the boundary layer in association with deep convection in a midlatitude case. The Arakawa-Schubert (1974) cumulus parameterization scheme was tested semi-prognostically against both the model result and the observation in a tropical rain event that occurred in GATE. The Cloud heating and drying effects predicted by the Arakawa-Schubert scheme were found to agree well both with the observation and the model result. However, it was also found that the Arakawa-Schuber scheme underestimates both condensation and evaporation rates substantially. An inclusion of the downdraft effect, as formulated by Johnson (1976), appears to alleviate this deficiency. This downdraft effect may be important in predicting the behavior of the boundary accurately.

Ogura, Y.↗
DOE OSTI2025

Designing robust energy policy packages under deep uncertainty: A multi-metric decision support framework

The complexity of transitioning to sustainable energy systems requires policy frameworks capable of balancing multiple objectives while addressing deep uncertainty. However, existing approaches often lack systematic methods to identify combinations of policy levers that remain effective across a wide range of uncertain futures. This paper presents a novel decision support framework that guides the selection of robust policy packages based on their performance across multiple objectives under uncertainty. Our method leverages a large ensemble of scenarios and applies scenario discovery techniques to identify influential policy levers. Here, we introduce new indicators to assess the robustness of policies by evaluating their ability to mitigate adverse outcomes across metrics. These indicators support an iterative process to build a robust policy package. Finally, we map the technological and energy pathways associated with the robust policy package by leveraging an energy system optimization model. We illustrate the application of this framework to the Spanish energy system, providing insights into how specific combinations of policy levers shape decarbonization pathways under uncertainty.

Decision-support method↗
DOE OSTI2022

Dimensionally Reduced Model for Rapid and Accurate Prediction of Gas Saturation, Pressure, and Brine Production in a CO 2 Storage Application: Case Study Using the SACROC Field as Part of SMART Task 5

This technical report presents work conducted by the sub surface analysis team of the Strategic Systems Analysis & Engineering group at NETL for Task 5 of SMART Phase 1. This study involved the development of deep learning models for CO 2 geologic storage that are capable of accurate prediction of spatio-temporal outputs of CO 2 saturation, pressure, and brine production in three dimensional space over a storage operation's injection and post-injection timeframes. The model framework involves ensembling multi-layer encoder networks that provide dimesionality reduction of geologic inputs with fully connected long short-term memory (LSTM) neural networks that generate time-series prediction This approach offers a means to maximize training time efficiency, reduce computational memory burden, and minimize prediction turnaround.

54 ENVIRONMENTAL SCIENCES↗
NASA NTRS2004

A New Approach to Using a Cloud-Resolving Model to Study the Interactions between Clouds, Precipitation and Aerosols

Recently Grabowski and Khairoutdinov and Randall have proposed and demonstrated, the feasibility of using 2D CRMs (these did not include detailed land processes and used periodic lateral boundary conditions) as 'super-parameterizations or a multi-scale modeling framework MMF' for cloud processes within atmospheric generation circulation models (GCMs). In the MMF, a fine-resolution 2D CRM takes the place of the single-column parameterization that is used in conventional GCMs. An overview of this approach is given by Randall et al.. Since a CRM can explicitly simulate cloud processes at the natural space and time scales of clouddynamical processes, it's possible to compute statistics of cloud fields, including radiances and radar backscatters, directly simulated/produced by the CRM. A seed fund is available at NASA Goddard to build a MMF based on a 2D GCE model and the Goddard fvGCM. A prototype MMF will be developed at end of 2004. The major objectives are to test and evaluate its performance using the GCE model as the super parameterization and compare its results with other super parameterizations, conventional GCMs and observations. Recently, a detailed spectral-bin microphysical scheme was implemented into the Goddard Cumulus Ensemble (WE) model. The GCE model is a CRM and it has been improved and applied to study the precipitation processes in the past two decades. Atmospheric aerosols are also described using number density size-distribution functions. The model is tested by studying the evolution of deep tropical clouds in the west Pacific warm pool region and summertime convection over a mid-latitude continent with different concentrations of CCN: a low 'clean' concentration and a high 'dirty' concentration. The impact of atmospheric aerosol concentration on cloud and precipitation will be investigated. In addition, the similarities and differences between bulk- and spectral-bin microphysics will be presented,

Tao, Wei-Kuo↗
DOE OSTI2021

Disentangling Rotational Dynamics and Ordering Transitions in a System of Self-Organizing Protein Nanorods via Rotationally Invariant Latent Representations

The dynamics of complex ordering systems with active rotational degrees of freedom exemplified by protein self-assembly is explored using a machine learning workflow that combines deep learning-based semantic segmentation and rotationally invariant variational autoencoder-based analysis of orientation and shape evolution. The latter allows for disentanglement of the particle orientation from other degrees of freedom and compensates for lateral shifts. The disentangled representations in the latent space encode the rich spectrum of local transitions that can now be visualized and explored via continuous variables. The time dependence of ensemble averages allows insight into the time dynamics of the system and, in particular, illustrates the presence of the potential ordering transition. Finally, analysis of the latent variables along the single-particle trajectory allows tracing these parameters on a single-particle level. The proposed approach is expected to be universally applicable for the description of the imaging data in optical, scanning probe, and electron microscopy seeking to understand the dynamics of complex systems where rotations are a significant part of the process.

representation learning↗