Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Vector Potential”

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
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

Investigation of the background in coherent J/ψ production at the EIC

Understanding various fundamental properties of nucleons and nuclei is among the most important scientific goals at the upcoming Electron-Ion Collider (EIC). With the unprecedented opportunity provided by the next-generation machine, the EIC might provide definitive answers to many standing puzzles and open questions in modern nuclear physics. We investigate one of the golden measurements proposed at the EIC, which is to obtain the spatial gluon density distribution within a lead (Pb) nucleus. The proposed experimental process is the exclusive J/ψ vector-meson production off the Pb nucleus: e + Pb → e' + J/ψ + Pb'. The Fourier transformation of the momentum transfer |t| distribution of the coherent diffraction is the transverse gluon spatial distribution. In order to measure it, the experiment has to overcome an overwhelmingly large background arising from the incoherent diffractive production, where the nucleus Pb' mostly breaks up into fragments of particles in the far-forward direction close to the hadron-going beam rapidity. We systematically study the rejection of incoherent J/ψ production by vetoing products from these nuclear breakups—protons, neutrons, and photons—which is based on the BeAGLE event generator and the most up-to-date EIC Far-forward Interaction Region design. The achieved vetoing efficiency, the ratio between the numbers of vetoed events and total incoherent events, ranges from about 80% to 99% depending on |t|. Assuming a 5% smearing applied to the reconstructed |t| resolution in the Sar t re model, this vetoing efficiency can suppress the incoherent background to at least the first minimum of the coherent |t| distribution. Experimental and accelerator machine challenges as well as potential improvements are discussed herein.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2024

Exciting DeePMD: Learning excited-state energies, forces, and non-adiabatic couplings

We extend the DeePMD neural network architecture to predict electronic structure properties necessary to perform non-adiabatic dynamics simulations. While learning the excited state energies and forces follows a straightforward extension of the DeePMD approach for ground-state energies and forces, how to learn the map between the non-adiabatic coupling vectors (NACV) and the local chemical environment descriptors of DeePMD is less trivial. Most implementations of machine-learning-based non-adiabatic dynamics inherently approximate the NACVs, with an underlying assumption that the energy-difference-scaled NACVs are conservative fields. We overcome this approximation, implementing the method recently introduced by Richardson [J. Chem. Phys. 158, 011102 (2023)], which learns the symmetric dyad of the energy-difference-scaled NACV. Furthermore, the efficiency and accuracy of our neural network architecture are demonstrated through the example of the methaniminium cation CH 2 NH 2 + .

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2025

Evaluation of the 2022 West Nile virus forecasting challenge, USA

Abstract Background West Nile virus (WNV) is the most common cause of mosquito-borne disease in the continental USA, with an average of ~1200 severe, neuroinvasive cases reported annually from 2005 to 2021 (range 386–2873). Despite this burden, efforts to forecast WNV disease to inform public health measures to reduce disease incidence have had limited success. Here, we analyze forecasts submitted to the 2022 WNV Forecasting Challenge, a follow-up to the 2020 WNV Forecasting Challenge. Methods Forecasting teams submitted probabilistic forecasts of annual West Nile virus neuroinvasive disease (WNND) cases for each county in the continental USA for the 2022 WNV season. We assessed the skill of team-specific forecasts, baseline forecasts, and an ensemble created from team-specific forecasts. We then characterized the impact of model characteristics and county-specific contextual factors (e.g., population) on forecast skill. Results Ensemble forecasts for 2022 anticipated a season at or below median long-term WNND incidence for nearly all (> 99%) counties. More counties reported higher case numbers than anticipated by the ensemble forecast median, but national caseload (826) was well below the 10-year median (1386). Forecast skill was highest for the ensemble forecast, though the historical negative binomial baseline model and several team-submitted forecasts had similar forecast skill. Forecasts utilizing regression-based frameworks tended to have more skill than those that did not and models using climate, mosquito surveillance, demographic, or avian data had less skill than those that did not, potentially due to overfitting. County-contextual analysis showed strong relationships with the number of years that WNND had been reported and permutation entropy (historical variability). Evaluations based on weighted interval score and logarithmic scoring metrics produced similar results. Conclusions The relative success of the ensemble forecast, the best forecast for 2022, suggests potential gains in community ability to forecast WNV, an improvement from the 2020 Challenge. Similar to the previous challenge, however, our results indicate that skill was still limited with general underprediction despite a relative low incidence year. Potential opportunities for improvement include refining mechanistic approaches, integrating additional data sources, and considering different approaches for areas with and without previous cases. Graphical Abstract

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Using Machine Learning to Predict Cloud Turbulent Entrainment–Mixing Processes

Different turbulent entrainment–mixing mechanisms between clouds and environment are essential to cloud–related processes; however, accurate representation of entrainment–mixing in weather/climate models still poses a challenge. This study exploits the use of machine learning (ML) to address this challenge. Four ML (Light Gradient Boosting Machine [LGB], eXtreme Gradient Boosting, Random Forest, and Support Vector Regression) are examined and compared. It is found that LGB performs best, and thus is selected to understand the impact of entrainment–mixing on microphysics using simulation data from Explicit Mixing Parcel Model. Compared with traditional parameterizations, the trained LGB provides more accurate microphysical properties (number concentration and cloud droplet spectral dispersion). The partial dependences of predicted microphysics on features exhibit a strong alignment with physical mechanisms and expectations, as determined by the interpreting method, thus overcoming the limitations of the “black box” scheme. The underlying mechanisms are that the smaller number concentration and larger spectral dispersion correspond to more inhomogeneous entrainment–mixing. Specifically, number concentration after entrainment–mixing is positively correlated with adiabatic number concentration and liquid water content affected by entrainment–mixing, and inversely correlated with adiabatic volume mean radius. Spectral dispersion after entrainment–mixing is negatively correlated with liquid water content affected by entrainment–mixing, turbulent dissipation rate and relative humidity of entrained air. Sensitivity analysis further suggests that number concentration is mainly determined by cloud microphysical properties whereas spectral dispersion is influenced by both cloud microphysical properties and environmental variables. The results indicate that the LGB scheme has the potential to enhance the representation of entrainment–mixing in weather/climate models.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2020

Bacterial genome editing by coupling Cre-lox and CRISPR-Cas9 systems

The past decade has been a golden age for microbiology, marked by the discovery of an unprecedented increase in the number of novel bacterial species. Yet gaining biological knowledge of those organisms has not kept pace with sequencing efforts. To unlock this genetic potential there is an urgent need for generic (i.e. non-species specific) genetic toolboxes. Recently, we developed a method, termed chassis-independent recombinase-assisted genome engineering (CRAGE), enabling the integration and expression of large complex gene clusters directly into the chromosomes of diverse bacteria. Here we expand upon this technology by incorporating CRISPR-Cas9 allowing precise genome editing across multiple bacterial species. To do that we have developed a landing pad that carries one wild-type and two mutant lox sites to allow integration of foreign DNA at two locations through Cre-lox recombinase-mediated cassette exchange (RMCE). The first RMCE event is to integrate the Cas9 and the DNA repair protein genes RecET, and the second RMCE event enables the integration of customized sgRNA and a repair template. Following this workflow, we achieved precise genome editing in four different gammaproteobacterial species. We also show that the inserted landing pad and the entire editing machinery can be removed scarlessly after editing. We report here the construction of a single landing pad transposon and demonstrate its functionality across multiple species. The modular design of the landing pad and accessory vectors allows design and assembly of genome editing platforms for other organisms in a similar way. We believe this approach will greatly expand the list of bacteria amenable to genetic manipulation and provides the means to advance our understanding of the microbial world.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2021

Galaxy–galaxy lensing with the DES-CMASS catalogue: measurement and constraints on the galaxy-matter cross-correlation

ABSTRACT The DMASS sample is a photometric sample from the DES Year 1 data set designed to replicate the properties of the CMASS sample from BOSS, in support of a joint analysis of DES and BOSS beyond the small overlapping area. In this paper, we present the measurement of galaxy–galaxy lensing using the DMASS sample as gravitational lenses in the DES Y1 imaging data. We test a number of potential systematics that can bias the galaxy–galaxy lensing signal, including those from shear estimation, photometric redshifts, and observing conditions. After careful systematic tests, we obtain a highly significant detection of the galaxy–galaxy lensing signal, with total S/N = 25.7. With the measured signal, we assess the feasibility of using DMASS as gravitational lenses equivalent to CMASS, by estimating the galaxy-matter cross-correlation coefficient rcc. By jointly fitting the galaxy–galaxy lensing measurement with the galaxy clustering measurement from CMASS, we obtain $r_{\rm cc}=1.09^{+0.12}_{-0.11}$ for the scale cut of $4 \, h^{-1}{\rm \,\,Mpc}$ and $r_{\rm cc}=1.06^{+0.13}_{-0.12}$ for $12 \, h^{-1}{\rm \,\,Mpc}$ in fixed cosmology. By adding the angular galaxy clustering of DMASS, we obtain rcc = 1.06 ± 0.10 for the scale cut of $4 \, h^{-1}{\rm \,\,Mpc}$ and rcc = 1.03 ± 0.11 for $12 \, h^{-1}{\rm \,\,Mpc}$. The resulting values of rcc indicate that the lensing signal of DMASS is statistically consistent with the one that would have been measured if CMASS had populated the DES region within the given statistical uncertainty. The measurement of galaxy–galaxy lensing presented in this paper will serve as part of the data vector for the forthcoming cosmology analysis in preparation.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2026

Discovering novel therapeutic V H Hs for emerging viruses: perspectives from VEEV selection strategies

Introduction: Evolution or emergence of a new viral variant is a significant public health concern. Alphaviruses, such as Venezuelan equine encephalitis virus (VEEV), are mosquito-borne viruses which are becoming more prevalent due to expansion of vector habitats. Despite this, there are currently no antiviral therapies or FDA-approved vaccines available to treat or prevent VEEV infection. The increased prevalence of such viruses provides opportunities for novel variants to evolve. Key therapeutic molecules that could be developed against viral pathogens are recombinant antibodies or antibody fragments, such as the variable heavy domain of heavy chain antibodies (V H Hs). Methods: In vitro selections offer a promising pathway for identification of therapeutic antibodies, here we explored isolation of V H Hs using phage and yeast display methodology with three antigen formats 1) recombinant E2, 2) linear peptides of E2, selected based on molecular dynamics analysis, and 3) UV inactivated virus. Results: Here we report four novel “human” V H Hs which bind to the VEEV E2 protein selected using different strategies that include both computational and biochemical design of suitable antigens and whole virus selections. These V H Hs have distinct complementarity-determining regions (CDRs). Multiple VHHs bind to the VEEV viral particles in ELISAs, and we report the peptide epitope recognized by these V H Hs. Discussion: Though non-neutralizing, these V H Hs bind to and sequester VEEV viral particles preventing infection, demonstrating the potential of these V H Hs to perform viral “sponging” which represents a novel therapeutic approach. The selection strategies we report may have applications to further antibody developments against other viruses.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2022

Machine learning models for rat multigeneration reproductive toxicity prediction

Reproductive toxicity is one of the prominent endpoints in the risk assessment of environmental and industrial chemicals. Due to the complexity of the reproductive system, traditional reproductive toxicity testing in animals, especially guideline multigeneration reproductive toxicity studies, take a long time and are expensive. Therefore, machine learning, as a promising alternative approach, should be considered when evaluating the reproductive toxicity of chemicals. We curated rat multigeneration reproductive toxicity testing data of 275 chemicals from ToxRefDB (Toxicity Reference Database) and developed predictive models using seven machine learning algorithms (decision tree, decision forest, random forest, k-nearest neighbors, support vector machine, linear discriminant analysis, and logistic regression). A consensus model was built based on the seven individual models. An external validation set was curated from the COSMOS database and the literature. The performances of individual and consensus models were evaluated using 500 iterations of 5-fold cross-validations and the external validation data set. The balanced accuracy of the models ranged from 58% to 65% in the 5-fold cross-validations and 45%–61% in the external validations. Prediction confidence analysis was conducted to provide additional information for more appropriate applications of the developed models. The impact of our findings is in increasing confidence in machine learning models. We demonstrate the importance of using consensus models for harnessing the benefits of multiple machine learning models (i.e., using redundant systems to check validity of outcomes). While we continue to build upon the models to better characterize weak toxicants, there is current utility in saving resources by being able to screen out strong reproductive toxicants before investing in vivo testing. The modeling approach (machine learning models) is offered for assessing the rat multigeneration reproductive toxicity of chemicals. Our results suggest that machine learning may be a promising alternative approach to evaluate the potential reproductive toxicity of chemicals.

consensus model↗
DOE OSTI2026

Leveraging large language models to address data scarcity in machine learning for graphene synthesis

Machine learning in experimental materials science faces significant challenges due to the scarcity of data, which are costly and time-consuming to generate, particularly when relying on in-house experiments. Literature data mining offers a potential solution but introduces issues like mixed data quality, inconsistent formats, and non-uniform reporting of synthesis parameters, resulting in partially missing and heterogeneous features across the dataset. Here, we propose data imputation and feature engineering methods that employ pre-trained large language models (LLMs) to enhance machine learning performance on scarce, heterogeneous datasets, demonstrated on graphene CVD synthesis data and the ML-HydPARK hydrogen storage dataset. GPT models perform data imputation via tailored prompting and semantic normalization of inconsistently reported features through embeddings, for example, to harmonize the complex nomenclature of CVD substrates. Beyond yielding more diverse and richer feature representations than traditional methods such as K-nearest neighbors (KNN) and Multivariate Imputation by Chained Equations (MICE), LLM-based data imputation is evaluated against dataset characteristics and prompting strategies. We vary the level of autonomy granted to the LLM, from generic prompting that leverages pre-trained knowledge for autonomous data generation to data-informed prompting that constrains outputs using target-specific information, and demonstrate which level of autonomy yields superior imputation performance across datasets and feature types. The proposed data engineering methods markedly improve downstream performance; for example, in graphene layer number classification using a support vector machine (SVM), binary accuracy increases from 39% to 65% and ternary accuracy from 52% to 72%. Fine-tuning experiments on both datasets show that combining our proposed LLM-based data imputation and feature encoding methods with numerical machine learning predictors outperforms standalone fine-tuned LLM predictors in data-scarce settings. The proposed strategies emphasize data enhancement techniques rather than refining learning architectures or regularizing loss functions, offering a broadly applicable framework for improving machine learning performance on scarce, inhomogeneous datasets.

Chemical vapor deposition↗
DOE OSTI2025

Evaluating the gradients of localized diabatic state energies and couplings at minimum cost

We calculate the gradients of Boys diabatic state energies and couplings when the electronic vector space is generated by configuration interaction singles. Our approach follows the Lagrangian approach of Paz and Glover (rather than direct differentiation of the adiabatic-to-diabatic approaches that have been published previously). The result is that we achieve a dramatic increase in savings over previous approaches, and the present approach should be immediately useful to scientists focused on electronic relaxation, especially chemists studying electron transfer who wish to go beyond the Condon approximation. Here, a future extension to time-dependent density functional theory in the Tamm–Dancoff approximation is clear.

Chemical physics↗
DOE OSTI2026

Low-Lying Excited States of Linear All- Trans Polyenes: Insights from Analytic Gradient and Nonadiabatic Coupling Calculations Based on Multireference Configuration Interaction

Polyenes serve as a rigorous test for theoretical models and electronic structure methods, playing a key role in advancing computational and theoretical chemistry. Here, we present a high-level theoretical investigation of linear, all-trans polyenes using energy gradients and nonadiabatic coupling vectors based on an MR-CISD wave function to describe electronic transitions involving the ground state (1 1 A g – ) and three low-lying excited states (2 1 A g – , 1 1 B u + , and 2 1 B u – ) of hexatriene, octatetraene, and decapentaene. This approach enables accurate evaluation of both adiabatic and vertical excitation and emission energies, yielding results in excellent agreement with experiment, as well as locating minima on the crossing seam between adiabatic states. Our results show that vertical excitation energies to the 1 1 B u + state are blue-shifted by 0.2–0.3 eV relative to the experimental absorption maximum, whereas the vertical emission energy from the 2 1 A g – state is red-shifted by ∼0.2 eV relative to the experimental emission maximum. Upon relaxation from the Franck–Condon geometry, the 2 1 A g – state stabilizes by around 1 eV, compared to 0.2–0.3 eV for the 1 1 B u + state. An analysis of the S 1 /S 0 crossing seam in hexatriene shows that its minimum involves asymmetric backbone deformations and provides an efficient channel for ultrafast internal conversion to the ground state, consistent with the absence of detectable fluorescence in this molecule. These results demonstrate the power of analytic gradients and nonadiabatic coupling vectors based on an MR-CISD wave function for accurately characterizing the electronic structure and photophysics of polyenes.

Excited states↗
DOE OSTI2021

Nonadiabatic couplings from a variational excited state method based on constrained DFT

Excited Costrained Density Functional Theory (XCDFT) is a variational excited state method that extends ground state DFT to the computation of low-lying excited states. It borrows much of the machinery of Constrained DFT (CDFT) with a crucial difference: the constraint imposes a population of one electron in the Hilbert space spanned by the virtuals of a reference ground state. Here, we present theory and implementation for evaluating nonadiabatic coupling vectors (NACVs) between the first excited state computed with XCDFT and the ground state. Our NACVs are computed analytically using density functional perturbation theory with a formalism that is general enough that could be applied to CDFT diabatic states. We showcase the new method with pilot NACV calculations for the conical intersection in H 3 , the avoided crossing in selenoacrolein, and the NACV magnitudes in azobenzene. Despite complications from the nonorthogonality of the wavefunctions, XCDFT’s energy surfaces and NACVs reproduce benchmark values and respect known sum rules within a reasonable degree. This shows that XCDFT is a viable method for nonadiabatic dynamics simulations.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2020

High Rate of Non-Human Feeding by Aedes aegypti Reduces Zika Virus Transmission in South Texas

Mosquito-borne viruses are emerging or re-emerging globally, afflicting millions of people around the world. Aedes aegypti, the yellow fever mosquito, is the principal vector of dengue, Zika, and chikungunya viruses, and has well-established populations across tropical and subtropical urban areas of the Americas, including the southern United States. While intense arboviral epidemics have occurred in Mexico and further south in the Americas, local transmission in the United States has been minimal. Here, we study Ae. aegypti and Culex quinquefasciatus host feeding patterns and vertebrate host communities in residential environments of South Texas to identify host-utilization relative to availability. Only 31% of Ae. aegypti blood meals were derived from humans, while 50% were from dogs and 19% from other wild and domestic animals. In Cx. quinquefasciatus, 67% of blood meals were derived from chicken, 22% came from dogs, 9% from various wild avian species, and 2% from other mammals including one human, one cat, and one pig. We developed a model for the reproductive number, R0, for Zika virus (ZIKV) in South Texas relative to northern Mexico using human disease data from Tamaulipas, Mexico. We show that ZIKV R0 in South Texas communities could be greater than one if the risk of human exposure to Ae. aegypti bites in these communities is at least 60% that of Northern Mexico communities. The high utilization of non-human vertebrates and low risk of human exposure in South Texas diminishes the outbreak potential for human-amplified urban arboviruses transmitted by Ae. aegypti.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2024

Energy Requirements for Integration of Nuclear Reactors with Iron and Steel Plants

This report identifies energy needs of heavy energy users within the domestic iron and steel industry and suggests solutions for integrating nuclear energy. The iron and steel industry, composed of several types of plants which perform different processes with varied energy demands and vectors, is a heavy consumer of electric power and fossil fuels including coke and natural gas. Almost all major process temperatures exceed the temperatures of direct heat available from advanced reactors, and so electricity and hydrogen were considered instead. Reference units were adopted and estimated energy demands computed for the blast furnace (BF), direct reduced iron (DRI) unit, electric arc furnace (EAF), and reheat furnaces. By utilizing production capacity data from industry reports, the ranges of power demands were estimated, including for hydrogen production by high temperature steam electrolysis (HTSE). For the EAF and DRI unit, more detailed integration studies with thermodynamic modeling were also conducted and determined a possible solution with a specific reactor design and number of modules. Furthermore, because many unit processes are co-located, entire plants were considered by adding the energy demands of the unit processes to form four hypothetical reference plants. The range of power needs for the reference plants is compatible with multi-unit banks of microreactors at the low end, and would create a need for multiple larger-capacity SMRs at the high end (1 GWe plus 0.14 GWt). Although the overall power need at the high end is well-matched with one present-day large reactor offering (1.1 GWe), redundancy considerations may require a minimum of two reactors, potentially eliminating the single large reactor from consideration. Auxiliary or house loads would increase the reference plant estimates. Finally, the report provides total estimated energy needs under integration of all U.S. units of each process (BF, DRI, EAF, and reheat furnaces), representing a national potential for nuclear energy in the industry. U.S. iron and steel plants may be candidates for integration with nuclear reactors via electricity and hydrogen, and many sites have energy requirements that correspond well to the capacities of several advanced nuclear power designs.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2021

Importance of Engineered and Learned Molecular Representations in Predicting Organic Reactivity, Selectivity, and Chemical Properties

Machine-readable chemical structure representations are foundational in all attempts to harness machine learning for the prediction of reactivities, selectivities, and chemical properties directly from molecular structure. The featurization of discrete chemical structures into a continuous vector space is a critical phase undertaken before model selection, and the development of new ways to quantitatively encode molecules is an active area of research. Here, we highlight the application and suitability of different representations, from expert-guided “engineered” descriptors to automatically “learned” features, in different prediction tasks relevant to organic and organometallic chemistry, where differing amounts of training data are available. These tasks include statistical models of stereo- and enantioselectivity, thermochemistry, and kinetics developed using experimental and quantum chemical data. The use of expert-guided molecular descriptors provides an opportunity to incorporate chemical knowledge, domain expertise, and physical constraints into statistical modeling. In applications to stereoselective organic and organometallic catalysis, where data sets may be relatively small and 3D-geometries and conformations play an important role, mechanistically informed features can be used successfully to obtain predictive statistical models that are also chemically interpretable. We provide an overview of several recent applications of this approach to obtain quantitative models for reactivity and selectivity, where topological descriptors, quantum mechanical calculations of electronic and steric properties, along with conformational ensembles, all feature as essential ingredients of the molecular representations used. Alternatively, more flexible, general-purpose molecular representations such as attributed molecular graphs can be used with machine learning approaches to learn the complex relationship between a structure and prediction target. This approach has the potential to out-perform more traditional representation methods such as “hand-crafted” molecular descriptors, particularly as data set sizes grow. One area where this is particularly relevant is in the use of large sets of quantum mechanical data to train quantitative structure–property relationships. A general approach toward curating useful data sets and training highly accurate graph neural network models is discussed in the context of organic bond dissociation enthalpies, where this strategy outperforms regression using precomputed descriptors. Finally, we describe how graph neural network predictions can be incorporated into mechanistically informed statistical models of chemical reactivity and selectivity. Once trained, this approach avoids the expensive computational overhead associated with quantum mechanical calculations, while maintaining chemical interpretability. We illustrate examples for which fast predictions of bond dissociation enthalpy and of the identities of radicals formed through cleavage of a molecule’s weakest bond are used in simple physical models of site-selectivity and reactivity.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

Self-Supervised Cloud Classification

Abstract Low-level marine clouds play a pivotal role in Earth’s weather and climate through their interactions with radiation, heat and moisture transport, and the hydrological cycle. These interactions depend on a range of dynamical and microphysical processes that result in a broad diversity of cloud types and spatial structures, and a comprehensive understanding of cloud morphology is critical for continued improvement of our atmospheric modeling and prediction capabilities moving forward. Deep learning has recently accelerated our ability to study clouds using satellite remote sensing, and machine learning classifiers have enabled detailed studies of cloud morphology. A major limitation of deep learning approaches to this problem, however, is the large number of hand-labeled samples that are required for training. This work applies a recently developed self-supervised learning scheme to train a deep convolutional neural network (CNN) to map marine cloud imagery to vector embeddings that capture information about mesoscale cloud morphology and can be used for satellite image classification. The model is evaluated against existing cloud classification datasets and several use cases are demonstrated, including training cloud classifiers with very few labeled samples, interrogation of the CNN’s learned internal feature representations, cross-instrument application, and resilience against sensor calibration drift and changing scene brightness. The self-supervised approach learns meaningful internal representations of cloud structures and achieves comparable classification accuracy to supervised deep learning methods without the expense of creating large hand-annotated training datasets. Significance Statement Marine clouds heavily influence Earth’s weather and climate, and improved understanding of marine clouds is required to improve our atmospheric modeling capabilities and physical understanding of the atmosphere. Recently, deep learning has emerged as a powerful research tool that can be used to identify and study specific marine cloud types in the vast number of images collected by Earth-observing satellites. While powerful, these approaches require hand-labeling of training data, which is prohibitively time intensive. This study evaluates a recently developed self-supervised deep learning method that does not require human-labeled training data for processing images of clouds. We show that the trained algorithm performs competitively with algorithms trained on hand-labeled data for image classification tasks. We also discuss potential downstream uses and demonstrate some exciting features of the approach including application to multiple satellite instruments, resilience against changing image brightness, and its learned internal representations of cloud types. The self-supervised technique removes one of the major hurdles for applying deep learning to very large atmospheric datasets.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2026

Isotopic discrepancy in microscopic fusion of 16 O + 40,48 Ca

We present a detailed comparison of fusion in 16 O + 40,48 Ca within a fully microscopic time-dependent Hartree-Fock framework over sub-, near-, and above-barrier energies. While 16 O + 40 Ca fusion cross sections are reproduced within the available experimental energy range, calculations for the neutron-rich 16 O + 48 Ca system systematically overestimate the measured near-barrier cross sections. The discrepancy persists across the tested energy density functionals, indicating that it is not tied to a particular parametrization. Analysis of the dynamical interaction potentials reveals an outward shift and softening of the barrier region at the lowest collision energies. These results are then compared to a frozen density description where the barrier modifications are absent and the predicted cross sections are brought closer to experiment. Here, this demonstrates that the observed discrepancy is closely linked to dynamical surface rearrangement and isovector effects, and highlights the sensitivity of low-energy fusion to both the structure and dynamics of neutron-rich nuclei. This suggests that limitations in the isovector sector of conventional Skyrme EDFs in describing the structure of 48 Ca, as highlighted by CREX, are also manifested in the fusion dynamics of the 16 O + 48 Ca system.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2021

Combinations of Single Chain Variable Fragments From HIV Broadly Neutralizing Antibodies Demonstrate High Potency and Breadth

Broadly neutralizing antibodies (bNAbs) are currently being assessed in clinical trials for their ability to prevent HIV infection. Single chain variable fragments (scFv) of bNAbs have advantages over full antibodies as their smaller size permits improved diffusion into mucosal tissues and facilitates vector-driven gene expression. We have previously shown that scFv of bNAbs individually retain significant breadth and potency. Here we tested combinations of five scFv derived from bNAbs CAP256-VRC26.25 (V2-apex), PGT121 (N332-supersite), 3BNC117 (CD4bs), 8ANC195 (gp120-gp41 interface) and 10E8v4 (MPER). Either two or three scFv were combined in equimolar amounts and tested in the TZM-bl neutralization assay against a multiclade panel of 17 viruses. Experimental IC 50 and IC 80 data were compared to predicted neutralization titers based on single scFv titers using the Loewe additive and the Bliss-Hill model. Like full-sized antibodies, combinations of scFv showed significantly improved potency and breadth compared to single scFv. Combinations of two or three scFv generally followed an independent action model for breadth and potency with no significant synergy or antagonism observed overall although some exceptions were noted. The Loewe model underestimated potency for some dual and triple combinations while the Bliss-Hill model was better at predicting IC 80 titers of triple combinations. Given this, we used the Bliss-Hill model to predict the coverage of scFv against a 45-virus panel at concentrations that correlated with protection in the AMP trials. Using IC 80 titers and concentrations of 1μg/mL, there was 93% coverage for one dual scFv combination (3BNC117+10E8v4), and 96% coverage for two of the triple combinations (CAP256.25+3BNC117+10E8v4 and PGT121+3BNC117+10E8v4). Combinations of scFv, therefore, show significantly improved breadth and potency over individual scFv and given their size advantage, have potential for use in passive immunization.

60 APPLIED LIFE SCIENCES↗