Search NASA⌕ Search

SEARCH · Search NASA

Results for “Model interpretation”

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 181 records · Page 10

Identification of Prompt Proton Emission in $N = Z - 1$ 61 Ga: Isospin Symmetry at the Limit of Nuclear Binding

Excited states in the proton drip line nucleus 61 Ga were populated via the fusion-evaporation reaction 24 Mg ⁢( 40 Ca, 𝑝⁢2⁢𝑛)⁢ 61 Ga. The experimental setup at Argonne National Laboratory comprised a novel combination of the Gammasphere array with two CD-shaped double-sided Si-strip detectors inside the Microball CsI(Tl) charged-particle detection array, as well as the Neutron-Shell liquid scintillators and the Fragment Mass Analyzer. Owing to the setup’s unprecedented in-beam proton spectroscopy and tracking capabilities, a coincidence between a 957.6(5)-keV 𝛾 ray and a 1.876(24)-MeV proton line was observed, which identifies the quasibound proton 𝜋⁢𝑔 9/2 single-particle state in 61 Ga at 𝐸 𝑥 = 2150⁢(34) keV. This probes isospin symmetry at the limit of nuclear binding by providing a unique challenge for the shell-model interpretation of mirror nuclei beyond doubly magic 56 Ni .

Hrabar, Yuliia [Lund Univ. (Sweden)] (ORCID:000000↗

Lifetime measurements and shape coexistence in {sup 144}Dy

The known level scheme of {sup 144}Dy has been extended and lifetime measurements have been made with the recoil-distance Doppler-shift method. Reduced transition probabilities and deformations have been determined for four low-lying transitions. These states form part of the first observed band crossing, giving information on the change in nuclear deformation resulting from the rearrangement of h{sub 11/2} protons in the nucleus. Two bands built upon excited 10{sup +} states have been assigned pi(h{sub 11/2}){sup 2} prolate and nu(h{sub 11/2}){sup -2} oblate configurations with tau=12(2)ps and 0.01<tau < or approx. 16ns, respectively. These long lifetimes are reasoned to be a result of shape coexistence at low energy and moderate spin. A known four-quasiparticle dipole band has been extended to higher spin and lifetime measurements suggest a long-lived bandhead state. In this case, the excited states in the band may be consistent with a shears model interpretation of a magnetic dipole rotor. However, the measured B(M1)/B(E2) branching ratios reveal a larger than expected deformed rotational component compared with that in the analogous band in the lower mass isotone {sup 142}Gd.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Probabilistic neural networks for fluid flow surrogate modeling and data recovery

We consider the use of probabilistic neural networks for fluid flow surrogate modeling and data recovery. This framework is constructed by assuming that the target variables are sampled from a Gaussian distribution conditioned on the inputs. Consequently, the overall formulation sets up a procedure to predict the hyperparameters of this distribution which are then used to compute an objective function given training data. We demonstrate that this framework has the ability to provide for prediction confidence intervals based on the assumption of a probabilistic posterior, given an appropriate model architecture and adequate training data. The applicability of the present framework to cases with noisy measurements and limited observations is also assessed. To demonstrate the capabilities of this framework, we consider canonical regression problems of fluid dynamics from the viewpoint of reduced-order modeling and spatial data recovery for four canonical data sets. The examples considered in this study arise from (i) the shallow-water equations, (ii) a two-dimensional cylinder flow, (iii) the wake of a NACA0012 airfoil with a Gurney flap, and (iv) the NOAA sea surface temperature data set. Furthermore, the present results indicate that the probabilistic neural network not only produces a machine-learning-based fluid flow surrogate model but also systematically quantifies the uncertainty therein to assist with model interpretability.

42 ENGINEERING↗

A priori precision estimation for neutron triples counting

The nondestructive assay of Plutonium bearing items for criticality, safety, security, safeguards, inventory balance, process control, waste management and compliance is often undertaken using correlated neutron counting. In particular Multiplicity Shift Register analysis allows one to extract autocorrelation parameters from the pulse train which can, within the framework of a simple interpretational model, be related to the effective {sup 240}Pu spontaneous fission mass present. The effective {sup 240}Pu mass is a weighted sum of the {sup 238}Pu, {sup 240}Pu and {sup 242}Pu masses so if the relative isotopic composition of the Pu can be established from the measured {sup 240}Pu effective mass one can estimate the total Pu mass and also the masses of the individual isotopes, example the fissile species {sup 239}Pu and {sup 241}Pu. In multiplicity counting three counting rates are obtained. These are the Singles, Doubles and Triples rates. The Singles rate is just the gross, totals or trigger rate. The Doubles and Triples rates are calculated from factorial moments of the observed signal triggered neutron multiplicity distributions following spontaneous fission in the item and can be thought of as the rate of observed coincident pairs and coincident triplets on the pulse train. Coincident events come about because the spontaneous fission and induced fission chains taking place in the item result in bursts of neutrons. These remain time correlated during the detection process and so retain information, through the burst size distribution, about the Pu content. In designing and assessing the performance of a detector system to meet a given goal it is necessary to make a priori estimates of the counting precision for all three kinds of rates. This is non-trivial because the counting does not obey the familiar rules of a Poissonian counting experiment because the pulse train has time correlated events on it and the train is sampled by event triggered gates that may overlap. For Singles and Doubles simple approximate analytical empirical rules for how to estimate the variance have been developed guided by theory and refined by experiment. However, for Triples no equivalent rules have been put forward and tested until now. In this work we propose an analytical expression, the CSH relation, for the variance on the Triples count and exercise it against experimental data gathered for Pu items measured in the Los Alamos National Laboratory's Epithermal Neutron Multiplicity Counter (ENMC). Preliminary results are encouraging and reasonable agreement with observation, considered fit for scoping studies, is obtained. We have also looked at the behavior using Monte Carlo simulations. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

From Data to Insights: A Covariate Analysis of the IARPA BRIAR Dataset for Multimodal Biometric Recognition Algorithms at Altitude and Range

This paper examines covariate effects on fused whole body biometrics performance in the IARPA BRIAR dataset, specifically focusing on UAV platforms, elevated positions, and distances up to 1000 meters. The dataset includes outdoor videos compared with indoor images and controlled gait recordings. Normalized raw fusion scores relate directly to predicted false accept rates (FAR), offering an intuitive means for interpreting model results. A linear model is developed to predict biometric algorithm scores, analyzing their performance to identify the most influential covariates on accuracy at altitude and range. Weather factors like temperature, wind speed, solar loading, and turbulence are also investigated in this analysis. The study found that resolution and camera distance best predicted accuracy and findings can guide future research and development efforts in long-range/elevated/UAV biometrics and support the creation of more reliable and robust systems for national security and other critical domains.

Bolme, David↗

Effect of progressively increasing lithium conditioning on edge transport and stability in high triangularity NSTX H-modes

A sequence of H-mode discharges with increasing levels of pre-discharge lithium evaporation (dose) was conducted in high triangularity and elongation boundary shape in NSTX. Energy confinement increased, and recycling decreased with increasing lithium dose, similar to a previous lithium dose scan in medium triangularity and elongation plasmas. Data-constrained SOLPS interpretive modeling quantified the edge transport change: the electron particle diffusivity decreased by 10-30x. The electron thermal diffusivity decreased by 4x just inside the top of the pedestal, but increased by up to 5x very near the separatrix. These results provide a baseline expectation for lithium benefits in NSTX-U, which is optimized for a boundary shape similar to the one in this experiment.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

DynamicME: dynamic simulation and refinement of integrated models of metabolism and protein expression

Background Genome-scale models of metabolism and macromolecular expression (ME models) enable systems-level computation of proteome allocation coupled to metabolic phenotype. Results We develop DynamicME, an algorithm enabling time-course simulation of cell metabolism and protein expression. DynamicME correctly predicted the substrate utilization hierarchy on a mixed carbon substrate medium. We also found good agreement between predicted and measured time-course expression profiles. ME models involve considerably more parameters than metabolic models (M models). We thus generate an ensemble of models (each model having its rate constants perturbed), and then analyze the models by identifying archetypal time-course metabolite concentration profiles. Furthermore, we use a metaheuristic optimization method to calibrate ME model parameters using time-course measurements such as from a (fed-) batch culture. Finally, we show that constraints on protein concentration dynamics (“inertia”) alter the metabolic response to environmental fluctuations, including increased substrate-level phosphorylation and lowered oxidative phosphorylation. Conclusions Overall, DynamicME provides a novel method for understanding proteome allocation and metabolism under complex and transient environments, and to utilize time-course cell culture data for model-based interpretation or model refinement.

Mathematical & Computational Biology↗

Machine learning of factors for improving oyster hatchery production

Oyster aquaculture and restoration in the Chesapeake Bay are vital, yet hatcheries frequently struggle with inconsistent larval growth and sudden mass mortality events. Unpredictable disruptions in larval production cause large economic losses, represent a perceived risk to growers, and impede industry expansion. To better understand associations between production yield and its potential predictors, we applied machine learning (random forest, and neural network) and statistical (generalized additive model) models to a comprehensive dataset of environmental, water quality, and operational parameters from a Maryland oyster hatchery, aiming to identify key yield predictors and develop a robust forecasting tool. We used recursive Boruta algorithm for variable selection, pinpointing critical predictors, and employed cross-validation to fine-tune model settings. Shapley value analysis offered crucial insights into model interpretations, highlighting week number, Normalized Difference Vegetation Index, salinity, turbidity, and fecundity as primary drivers of yield variability. For low-yield cases, salinity-related variables were particularly important. Our findings provide an early warning system for potential production downturns, empowering hatchery operators to make data-driven decisions for optimizing water conditions, feeding schedules, and broodstock management. By boosting predictability and efficiency, this research directly supports economic stability of the oyster industry and ecological health of the Chesapeake Bay.

Vishwakarma, Srishti [Oak Ridge National Laborator↗

Search for squarks and gluinos using final states with jets and missing transverse momentum with the ATLAS detector in s = 7 TeV proton-proton collisions

CERN-LHC. Study of events having final states with no leptons, jets and missing transverse momentum, and with no reconstructed electrons or muons, in proton-proton collisions at a centre-of-mass energy of 7 TeV. Data are selected with various cuts and criteria to enhance different physics regions as described in the text of the paper. Exclusion limits on gluino and squark masses in SUSY models are presented for a simplified model (described in the paper) together with limits on m_0 and m_1/2 in MSUGRA/CMSSM models. The data sample has a total integrated luminosity of 1.04 FB-1 and shows no excess above the Standard Model background. Additional model interpretations are given in the Conference Note.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Linear viscoelastic measurements of commerical therapy putties for design and intuition

This dataset complements the publication Mapping Linear Viscoelasticity for Design and Tactile Intuition (Corman & Ewoldt, 2019). It contains rigorous shear rheometry measurements for six commercial silicone putties, including creep recovery, SAOS, LAOS, and stress relaxation. Additional, unpublished data was generated as part of a graduate rheology course taught by Randy Ewoldt at University of Illinois Urbana-Champaign. Together, these datasets provide a comprehensive example of linear viscoelasticity of a commercial material. The collection is intended both as a pedagogical resource for developing physical intuition in linear viscoelasticity and as a reference dataset for data interpretation, model fitting, and comparative rheological analysis.

C. Marsh, Maxwell↗

Modification Of The Electron Energy Distribution Function During Lithium Experiments On The National Spherical Torus Experiment

The National Spherical Torus Experiment (NSTX) has recently studied the use of a liquid lithium divertor (LLD). Divertor Langmuir probes have also been installed for making measurements of the local plasma conditions. A non-local probe interpretation method is used to supplement the classical probe interpretation and obtain measurements of the electron energy distribution function (EEDF) which show the occurrence of a hot-electron component. Analysis is made of two discharges within a sequence that exhibited changes in plasma fueling efficiency. It is found that the local electron temperature increases and that this increase is most strongly correlated with the energy contained within the hot-electron population. Preliminary interpretative modeling indicates that kinetic effects are likely in the NSTX.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Phase changes in fluoropolymer binders

Motivation: Understanding fluoropolymer binder structure and aging response continues to be an important goal for the prediction and assessment of polymer bonded explosive compositions within the stockpile. As part of ongoing efforts within NNM, we are developing new methods to quantify crystallization and chemical degradation within aged Fluoropolymer materials. Significance: Time, temperature and radiation are known to alter the mechanical properties of fluoropolymers as crystalline domains nucleate and grow within the material. However, current methods do not accurately quantify and assess the degree or rapidity of these physical aging processes. For this reason, we are developing methods that directly measure crystalline domain, composition, size and shape and cross-correlating them with traditional techniques. Progress: Most recently, we thermally aged fluoropolymer binder, FK800, for various durations up to 23 days at several temperatures including: 25C (red), 35C (green), and 60C (blue). We used wide angle X-ray scattering (WAXS) to observe when crystalline peaks emerged, signaling a phase transition from amorphous to partially crystalline. Some of these data are summarized in Figure 1. As expected, samples held at 25C (red), which is below the glass transition temperature at 28C, remain amorphous – the WAXS data (Figure 1b) lacks sharp peaks indicative of crystallinity and the ultra-small angle X-ray scattering (USAXS) also has a smooth background without peaks indicative of structure (Figure 1a). By contrast, samples held at 35C and 60C have new peaks in both the USAXS and WAXS data. The structure is clearer after subtracting off the amorphous background from the WAXS data (Figure 1c). These data show that samples aged at 60C are more crystalline as determined by the more defined peaks and also have a larger lattice constant as seen by the shift in peak position as compared to the 35C samples. The corresponding USAXS data (Figure 1a) indicate that the size of the crystals was larger at 60C than at 35C as seen by the shift in the knee position to lower q-values. Interestingly, the shape of the scatter curve also changed, which indicates a change in morphology. Atomic force microscopy (AFM) was used to image the fluoropolymer (Figure 1c-e). At 35C, ~50nm long stacked crystal plates are observed. High resolution imaging shows that the stacks consist of 2-6 lamellae that are spaced approximately 3nm apart. At 60C, the individual lamella appear both thicker and longer and are intertwined to fill the surface. Morphological data such as these images help to interpret & model volume averaging techniques like X-ray scattering and NMR. We also expect the shift from isolated islands to intertwined clusters to impact mechanical properties.

36 MATERIALS SCIENCE↗

Sandia-UT Academic Alliance Project Summary

This project seeks to leverage various hyperspectral tensor products for the purposes of target classification/detection/prediction. In addition to hyperspectral, these products may be images, time series, geometries, or other modalities. The scenarios in which the targets of interest must be identified are typically from remote sensing platforms such as satellites. As such, there are numerous real-world constraints that drive algorithmic formulation. Cost, complexity, and feasibility of the algorithm should all be considered. Targets of interest are exceedingly rare, and collecting many data samples is prohibitively expensive. Furthermore, model interpretability is paramount due to the application space. The goal of this project is to develop a constrained supervised tensor factorization framework for use on hyperspectral data products. Supervised tensor factorizations already exist in the literature, although they have not seen widespread adoption in the remote sensing domain. The novelty of this project will be the formulation and inclusion of constraints that take into account mission considerations and physics based limits to learn a factorization that is both physically interpretable and mission deployable. This will represent a new contribution to the field of remote sensing for performing supervised learning tasks with hyperspectral data.

97 MATHEMATICS AND COMPUTING↗

Ex-Situ Surface Characterization Studies and Boundary Plasma Diagnostic Development for DIII-D (Final Report)

This report details the accomplishments for award DE-SC0016318, which sponsored collaborative research activities between the University of Tennessee-Knoxville and the DIII-D experiment at General Atomics, led by the PI (Donovan), which officially began on August 1, 2016 and ended July 31, 2020. Though the official start date for the award was August 1, 2016, the PI had already initiated collaborative activities with DIII-D during 2015 supported by internal UTK start-up funds utilized by the PI. This prior work enabled UTK to have a substantial role in the June 2016 Metal Rings Campaign (MRC). The DOE funds then provided the opportunity to expand the UTK team with funding for students and a postdoc to perform ex-situ analysis on the wide array of samples exposed during the MRC and develop more sophisticated analysis tools and interpretive modeling techniques.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

NuGraph2: A Graph Neural Network for Neutrino Event Reconstruction

Neutrino experiments are set to probe some of the most important open questions in physics, from CP violation and the nature of dark matter. The technology of choice for many of these experiments is the liquid argon time projection chamber (LArTPC). In current LArTPC experiments, reconstruction performance often represents a limiting factor for the sensitivity. New developments are therefore needed to unlock the full potential of LArTPC experiments. NuGraph2 is a state of the art Graph Neural Network for reconstruction of data in LArTPC experiments. NuGraph2 utilizes a heterogeneous graph structure, with separate subgraphs of 2D nodes (hits in each plane) connected across planes via 3D nodes (space points). The model provides a consistent description of the neutrino interaction across all planes. NuGraph2 is a multi-purpose network, with a common message-passing attention engine connected to multiple decoders with different classification or regression tasks. These include the classification of detector hits according to the particle type that produced them (semantic segmentation) and the separation of hits from the neutrino interaction from hits due to noise or cosmic-ray background. Additional decoders are being developed, performing tasks such as the regression of the neutrino interaction vertex position. Performance results will be presented based on publicly available samples from MicroBooNE. These include both physics performance metrics, achieving 95% accuracy for semantic segmentation and 98% classification of neutrino hits, as well as computational metrics for training and for inference on CPU or GPU. The status of the NuGraph integration in the LArSoft software framework will be presented, as well as initial studies about model interpretability and injection of domain knowledge.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Synthetic Scientific Image Generation with VAE, GAN, and Diffusion Model Architectures

Generative AI (genAI) has emerged as a powerful tool for synthesizing diverse and complex image data, offering new possibilities for scientific imaging applications. This review presents a comprehensive comparative analysis of leading generative architectures, ranging from Variational Autoencoders (VAEs) to Generative Adversarial Networks (GANs) on through to Diffusion Models, in the context of scientific image synthesis. We examine each model's foundational principles, recent architectural advancements, and practical trade-offs. Our evaluation, conducted on domain-specific datasets including microCT scans of rocks and composite fibers, as well as high-resolution images of plant roots, integrates both quantitative metrics (SSIM, LPIPS, FID, CLIPScore) and expert-driven qualitative assessments. Results show that GANs, particularly StyleGAN, produce images with high perceptual quality and structural coherence. Diffusion-based models for inpainting and image variation, such as DALL-E 2, delivered high realism and semantic alignment but generally struggled in balancing visual fidelity with scientific accuracy. Importantly, our findings reveal limitations of standard quantitative metrics in capturing scientific relevance, underscoring the need for domain-expert validation. We conclude by discussing key challenges such as model interpretability, computational cost, and verification protocols, and discuss future directions where generative AI can drive innovation in data augmentation, simulation, and hypothesis generation in scientific research.

Generative Adversarial Networks↗

NuGraph2: A Graph Neural Network for Neutrino Event Reconstruction

Neutrino experiments are set to probe some of the most important open questions in physics, from CP violation and the nature of dark matter. The technology of choice for many of these experiments is the liquid argon time projection chamber (LArTPC). In current LArTPC experiments, reconstruction performance often represents a limiting factor for the sensitivity. New developments are therefore needed to unlock the full potential of LArTPC experiments. NuGraph2 is a state of the art Graph Neural Network for reconstruction of data in LArTPC experiments. NuGraph2 utilizes a heterogeneous graph structure, with separate subgraphs of 2D nodes (hits in each plane) connected across planes via 3D nodes (space points). The model provides a consistent description of the neutrino interaction across all planes. NuGraph2 is a multi-purpose network, with a common message-passing attention engine connected to multiple decoders with different classification or regression tasks. These include the classification of detector hits according to the particle type that produced them (semantic segmentation) and the separation of hits from the neutrino interaction from hits due to noise or cosmic-ray background. Additional decoders are being developed, performing tasks such as the regression of the neutrino interaction vertex position. Performance results will be presented based on publicly available samples from MicroBooNE. These include both physics performance metrics, achieving 95% accuracy for semantic segmentation and 98% classification of neutrino hits, as well as computational metrics for training and for inference on CPU or GPU. The status of the NuGraph integration in the LArSoft software framework will be presented, as well as initial studies about model interpretability and injection of domain knowledge.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

NuGraph2: A Graph Neural Network for Neutrino Event Reconstruction

Neutrino experiments are set to probe some of the most important open questions in physics, from CP violation and the nature of dark matter. The technology of choice for many of these experiments is the liquid argon time projection chamber (LArTPC). In current LArTPC experiments, reconstruction performance often represents a limiting factor for the sensitivity. New developments are therefore needed to unlock the full potential of LArTPC experiments. NuGraph2 is a state of the art Graph Neural Network for reconstruction of data in LArTPC experiments [https://arxiv.org/abs/2403.11872]. NuGraph2 utilizes a heterogeneous graph structure, with separate subgraphs of 2D nodes (hits in each plane) connected across planes via 3D nodes (space points). The model provides a consistent description of the neutrino interaction across all planes. NuGraph2 is a multi-purpose network, with a common message-passing attention engine connected to multiple decoders with different classification or regression tasks. These include the classification of detector hits according to the particle type that produced them (semantic segmentation) and the separation of hits from the neutrino interaction from hits due to noise or cosmic-ray background. Additional decoders are being developed, performing tasks such as the regression of the neutrino interaction vertex position. Performance results will be presented based on publicly available samples from MicroBooNE. These include both physics performance metrics, achieving 95% accuracy for semantic segmentation and 98% classification of neutrino hits, as well as computational metrics for training and for inference on CPU or GPU. The status of the NuGraph integration in the LArSoft software framework will be presented, as well as initial studies about model interpretability and injection of domain knowledge.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗