Search NASA⌕ Search

SEARCH · Search NASA

Results for “model interpretability”

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 1,405 records · Page 78

Application of theoretical models to active and passive remote sensing of saline ice

The random medium model is used to interpret the polarimetric active and passive measurements of saline ice. The ice layer is described as a host ice medium embedded with randomly distributed inhomogeneities, and the underlying sea water is considered as a homogeneous half-space. The scatterers in the ice layer are modeled with an ellipsoidal correlation function. The orientation of the scatterers is vertically aligned and azimuthally random. The strong permittivity fluctuation theory is employed to calculate the effective permittivity and the distorted Born approximation is used to obtain the polarimetric scattering coefficients. We also calculate the thermal emissions based on the reciprocity and energy conservation principles. The effects of the random roughness at the air-ice, and ice-water interfaces are accounted for by adding the surface scattering to the volume scattering return incoherently. The above theoretical model, which has been successfully applied to analyze the radar backscatter data of the first-year sea ice near Point Barrow, AK, is used to interpret the measurements performed in the CRRELEX program.

Han, H. C.↗

A Tale of Tails. Dark Matter Interpretations of the Fermi GeV Excess in Light of Background Model Systematics

Several groups have identified an extended excess of gamma rays over the modeled foreground and background emissions towards the Galactic center (GC) based on observations with the Fermi Large Area Telescope. The excess emission is compatible in morphology and spectrum with a telltale sign from dark matter (DM) annihilation. Here, we present a critical reassessment of DM interpretations of the GC signal in light of the foreground and background uncertainties that some of us recently outlaid in Calore et al. (2014). We also find that a much larger number of DM models fits the gamma-ray data than previously noted. In particular: (1) In the case of DM annihilation into b¯b, we find that even large DM masses up to m χ ≃74 GeV are allowed at p-value >0.05. (2) Surprisingly, annihilation into nonrelativistic hh gives a good fit to the data. (3) The inverse Compton emission from μ + μ - with m χ ~60–70 GeV can also account for the excess at higher latitudes, |b|>2°, both in its spectrum and morphology. We also present novel constraints on a large number of mixed annihilation channels, including cascade annihilation involving hidden sector mediators. Finally, we show that the current limits from dwarf spheroidal observations are not in tension with a DM interpretation when uncertainties on the DM halo profile are accounted for.

79 ASTRONOMY AND ASTROPHYSICS↗

Quantum filter of spin polarized states: Metal–dielectric–ferromagnetic/semiconductor device

Highlights: • Development of a new spintronics device. • Development of quantum spin polarized filters. • Development of theory of quantum spin polarized filter. - Abstract: Recently we proposed a model for the Quantum Spin-Polarized State Filter (QSPSF). The magnetic moments are transported selectively in this model, detached from the electric charge carriers. Thus, transfer of a spin-polarized state between two conductors was predicted in a system of two levels coupled by exchange interaction. The strength of the exchange interaction between the two conductive layers depends on the thickness of the dielectric layer separating them. External magnetic fields modulate spin-polarized state transfer, due to Zeeman level shift. Therefore, a linearly growing magnetic field generates a series of current peaks in a nearby coil. Thus, our spin-state filter should contain as least three nanolayers: (1) conductive or ferromagnetic; (2) dielectric; and (3) conductive or semiconductive. The spectrum of spin-polarized states generated by the filter device consists of a series of resonance peaks. In a simple case the number of lines equals S, the total spin angular momentum of discrete states in one of the coupled nanolayers. Presently we report spin-polarized state transport in metal–dielectric–ferromagnetic (MDF) and metal–dielectric–semiconductor (MDS) three-layer sandwich devices. The exchange-resonance spectra in such devices are quite specific, differing also from spectra observed earlier in other three-layer devices. The theoretical model is used to interpret the available experimental results. A detailed ab initio analysis of the magnetic-field dependence of the output magnetic moment averaged over the surface of the device was carried out. The model predicts the resonance structure of the signal, although at its present accuracy it cannot predict the positions of the spectral peaks.

36 MATERIALS SCIENCE↗

A Universal Machine Learning Model for Elemental Grain Boundary Energies

The grain boundary (GB) energy has a profound influence on the grain growth and properties of polycrystalline metals. Here, we show that the energy of a GB, normalized by the bulk cohesive energy, can be described purely by four geometric features. By machine learning on a large computed database of 361 small Σ (Σ<10) GBs of more than 50 metals, we develop a model that can predict the grain boundary energies to within a mean absolute error of 0.13 J m –2 . More importantly, this universal GB energy model can be extrapolated to the energies of high Σ GBs without loss in accuracy. These results highlight the importance of capturing fundamental scaling physics and domain knowledge in the design of interpretable, extrapolatable machine learning models for materials science.

36 MATERIALS SCIENCE↗

Measurable characteristics of lysozyme crystal growth

The behavior of protein crystal growth is estimated from measurements performed at both the microscopic and molecular levels. In the absence of solutal flow, it was determined that a model that balances the macromolecular flux toward the crystal surface with the flux of the crystal surface well characterizes crystal growth observed using microscopic methods. Namely, it was determined that the model provides accurate estimates for the crystal-growth velocities upon evaluation of crystal-growth measurements obtained in time. Growth velocities thus determined as a function of solution supersaturation were further interpreted using established deterministic models. From analyses of crystal-growth velocities, it was found that the mode of crystal growth varies with respect to increasing solution supersaturation, possibly owing to kinetic roughening. To verify further the hypothesis of kinetic roughening, crystal growth at the molecular level was examined using atomic force microscopy (AFM). From the AFM measurements, it was found that the magnitude of surface-height fluctuations, h(x), increases with increasing solution supersaturation. In contrast, the estimated characteristic length, xi, decreases rapidly upon increasing solution supersaturation. It was conjectured that the magnitude of both h(x) and xi could possibly determine the mode of crystal growth. Although the data precede any exact theory, the non-critical divergence of h(x) and xi with respect to increasing solution supersaturation was nevertheless preliminarily established. Moreover, approximate models to account for behavior of both h(x) and xi are also presented.

Muramidase/chemistry↗

Fingerprinting Interactions between Proteins and Ligands for Facilitating Machine Learning in Drug Discovery

Molecular recognition is fundamental in biology, underpinning intricate processes through specific protein–ligand interactions. This understanding is pivotal in drug discovery, yet traditional experimental methods face limitations in exploring the vast chemical space. Computational approaches, notably quantitative structure–activity/property relationship analysis, have gained prominence. Molecular fingerprints encode molecular structures and serve as property profiles, which are essential in drug discovery. While two-dimensional (2D) fingerprints are commonly used, three-dimensional (3D) structural interaction fingerprints offer enhanced structural features specific to target proteins. Machine learning models trained on interaction fingerprints enable precise binding prediction. Recent focus has shifted to structure-based predictive modeling, with machine-learning scoring functions excelling due to feature engineering guided by key interactions. Notably, 3D interaction fingerprints are gaining ground due to their robustness. Various structural interaction fingerprints have been developed and used in drug discovery, each with unique capabilities. This review recapitulates the developed structural interaction fingerprints and provides two case studies to illustrate the power of interaction fingerprint-driven machine learning. The first elucidates structure–activity relationships in β2 adrenoceptor ligands, demonstrating the ability to differentiate agonists and antagonists. The second employs a retrosynthesis-based pre-trained molecular representation to predict protein–ligand dissociation rates, offering insights into binding kinetics. Despite remarkable progress, challenges persist in interpreting complex machine learning models built on 3D fingerprints, emphasizing the need for strategies to make predictions interpretable. Binding site plasticity and induced fit effects pose additional complexities. Interaction fingerprints are promising but require continued research to harness their full potential.

3D structural interaction fingerprints↗

Integration of facility modeling capabilities for nuclear nonproliferation analysis

Developing automated methods for data collection and analysis that can facilitate nuclear nonproliferation assessment is an important research area with significant consequences for the effective global deployment of nuclear energy. Facility modeling that can integrate and interpret observations collected from monitored facilities in order to ascertain their functional details will be a critical element of these methods. Although improvements are continually sought, existing facility modeling tools can characterize all aspects of reactor operations and the majority of nuclear fuel cycle processing steps, and include algorithms for data processing and interpretation. Assessing nonproliferation status is challenging because observations can come from many sources, including local and remote sensors that monitor facility operations, as well as open sources that provide specific business information about the monitored facilities, and can be of many different types. Although many current facility models are capable of analyzing large amounts of information, they have not been integrated in an analyst-friendly manner. This paper addresses some of these facility modeling capabilities and illustrates how they could be integrated and utilized for nonproliferation analysis. The inverse problem of inferring facility conditions based on collected observations is described, along with a proposed architecture and computer framework for utilizing facility modeling tools. After considering a representative sampling of key facility modeling capabilities, the proposed integration framework is illustrated with several examples.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Co-optimized machine-learned manifold models for large eddy simulation of turbulent combustion

Many modeling approaches in large eddy simulation (LES) of turbulent combustion employ a projection of the thermochemical state onto a low-dimensional manifold within state space to reduce the number of transported variables and hence computational cost. Flamelet-generated manifolds (FGM) is an example of a well-established, physics-based approach, but increasingly, principal component analysis (PCA) is being used as a data-driven method for generating manifold models. For both approaches, the nonlinear relationship between the location on the predefined manifold and the outputs of interest, such as reaction rates, can be tabulated or encoded in a neural network. This work proposes a new approach for manifold modeling that extends these existing approaches. A modified neural network structure simultaneously encodes the definition of the manifold variables, the nonlinear mapping, and the subfilter closure for LES. This allows all three of these aspects of the model to be co-optimized, generating a model from any source of combustion thermochemical state data. The manifold parameterizing variables are constrained to be linear combinations of species, as in FGM and PCA-based models, to aid in interpretability and implementation. For LES, subfilter variances of the manifold variables are also included as inputs. Two types of a priori analysis are performed to evaluate the new approach. In the first, the model is trained on data from one-dimensional premixed flames. In this case, the approach recovers the behavior of flamelet-based manifold approaches, and in fact slightly improves performance by identifying an optimized progress variable. The approach is also applied to data from direct numerical simulations of spherical ignition kernels in isotropic turbulence. For any specified manifold dimensionality, the new approach provides substantially lower prediction errors than a PCA-based model developed from the same data set. Additionally, the LES formulation of the new approach can provide accurate predictions for filtered reaction rates across a variety of filter widths.

97 MATHEMATICS AND COMPUTING↗

Non-local Thermodynamic Equilibrium Stellar Spectroscopy with 1D and 〈3D〉 Models. II. Chemical Properties of the Galactic Metal-poor Disk and the Halo

From exploratory studies and theoretical expectations it is known that simplifying approximations in spectroscopic analysis (local thermodynamic equilibrium (LTE), 1D) lead to systematic biases of stellar parameters and abundances. These biases depend strongly on surface gravity, temperature and, in particular, for LTE versus non-LTE (NLTE), on metallicity of the stars. Here we analyze the [Mg/Fe] and [Fe/H] plane of a sample of 326 stars, comparing LTE and NLTE results obtained using 1D hydrostatic models and averaged 〈3D〉 models. We show that compared to the 〈3D〉 NLTE benchmark, the other three methods display increasing biases toward lower metallicities, resulting in false trends of [Mg/Fe] against [Fe/H], which have profound implications for interpretations by chemical evolution models. In our best 〈3D〉 NLTE model, the halo and disk stars show a clearer behavior in the [Mg/Fe]–[Fe/H] plane, from the knee in abundance space down to the lowest metallicities. Our sample has a large fraction of thick disk stars and this population extends down to at least [Fe/H] ∼ −1.6 dex, further than previously proven. The thick disk stars display a constant [Mg/Fe] ≈ 0.3 dex, with a small intrinsic dispersion in [Mg/Fe] that suggests that a fast SN Ia channel is not relevant for the disk formation. The halo stars reach higher [Mg/Fe] ratios and display a net trend of [Mg/Fe] at low metallicities, paired with a large dispersion in [Mg/Fe]. These indicate the diverse origin of halo stars from accreted low-mass systems to stochastic/inhomogeneous chemical evolution in the Galactic halo.

79 ASTRONOMY AND ASTROPHYSICS↗

Self-consistent models for triaxial galaxies with flat rotation curves - The disk case

We examine the possibility of constructing scale-free triaxial logarithmic potentials self-consistently, using Schwarzschild's linear programing method. In particular, we explore the limit of nonaxisymmetric disks. In this case it is possible to reduce the problem to the self-consistent reconstruction of the disk surface density on the unit circle, a considerably simpler problem than the usual 2D or 3D one. Models with surface densities of the form Sigma = (x exp n + (y/q) exp n) exp - 1/n with n = 2 or 4 are investigated. We show that the complicated shapes of the 'boxlet' orbit families (which replace the box orbit family found in potentials with smooth cores) limit the possibility of building self-consistent models, though elliptical disks of axis ratio above 0.7 and a restricted range of boxier models can be constructed. This result relies on using sufficiently fine bins, smaller than the 10 deg bins commonly used in 2D or 3D investigations. It also indicates the need for caution in interpreting N-body models of triaxial halos in which the core of the potential is numerically smoothed.

Kuijken, Konrad↗

TPSAS-NF1676L-10206-DND

NASA is at fore-front in developing active and passive remote sensing technologies and unique capabilities towards space-based observations for understanding the complexities and interactions among Earth system components. The world is facing significant environmental challenges and a robust, integrated, and flexible system of observations and models are needed for understanding the short-and long term impact on the Earth system. A fundamental challenge for the coming decade is to ensure that space-based observations, analyses, better interpretive understanding, enhanced predictive models, broadened international community participation, and improved means for information assimilation and disseminations are well coordinated to realize the full economic, societal, and security benefit of Earth science. This presentation will provide an overview of NASA?s future vision for Earth science missions for global observations and the challenges associated in applying them for societal benefit.

Upendra N Singh↗

NASA's Future Earth Science Missions for Global Observations

NASA is at fore-front in developing active and passive remote sensing technologies and unique capabilities towards space-based observations for understanding the complexities and interactions among Earth system components. The world is facing significant environmental challenges and a robust, integrated, and flexible system of observations and models are needed for understanding the short-and long term impact on the Earth system. A fundamental challenge for the coming decade is to ensure that space-based observations, analyses, better interpretive understanding, enhanced predictive models, broadened international community participation, and improved means for information assimilation and disseminations are well coordinated to realize the full economic, societal, and security benefit of Earth science. This presentation will provide an overview of enabling active and passive remote sensing technologies and techniques, NASA's future vision for Earth science missions for global observations, and the challenges associated in applying them for societal benefit.

Upendra N Singh↗

TPSAS-NF1676L-27491-DND

NASA is at fore-front in developing lidar technologies and unique active/passive remote sensing capabilities towards space-based observations for understanding the complexities and interactions among Earth system components. The world is facing significant environmental challenges and a robust, integrated, and flexible system of observations and models are needed for understanding the short-and long term impact on the Earth system. A fundamental challenge for the coming decade is to ensure that space-based observations, analyses, better interpretive understanding, enhanced predictive models, broadened international community participation, and improved means for information assimilation and disseminations are well coordinated to realize the full economic, societal, and security benefit of Earth science. This presentation will provide an overview of enabling lidar technologies and techniques from ground and space towards NASA's future vision for Earth science missions for global observations, and the challenges associated in applying them for societal benefit.

Upendra N Singh↗

Physiologically based pharmacokinetic toolkit to evaluate environmental exposures: Applications of the dioxin model to study real life exposures

Chlorinated dibenzo-p-dioxins (CDDs) are a series of mono- to octa-chlorinated homologous chemicals commonly referred to as polychlorinated dioxins. One of the most potent, well-known, and persistent member of this family is 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). As part of translational research to make computerized models accessible to health risk assessors, we present a Berkeley Madonna recoded version of the human physiologically based pharmacokinetic (PBPK) model used by the U.S. Environmental Protection Agency (EPA) in the recent dioxin assessment. This model incorporates CYP1A2 induction, which is an important metabolic vector that drives dioxin distribution in the human body, and it uses a variable elimination half-life that is body burden dependent. To evaluate the model accuracy, the recoded model predictions were compared with those of the original published model. The simulations performed with the recoded model matched well with those of the original model. The recoded model was then applied to available data sets of real life exposure studies. The recoded model can describe acute and chronic exposures and can be useful for interpreting human biomonitoring data as part of an overall dioxin and/or dioxin-like compounds risk assessment. - Highlights: • The best available dioxin PBPK model for interpreting human biomonitoring data is presented. • The original PBPK model was recoded from acslX to the Berkeley Madonna (BM) platform. • Comparisons were made of the accuracy of the recoded model with the original model. • The model is a useful addition to the ATSDR's BM based PBPK toolkit that supports risk assessors. • The application of the model to real-life exposure data sets is illustrated.

54 ENVIRONMENTAL SCIENCES↗

Search for resonant pair production of Higgs bosons in the $\textrm{b}\overline{\textrm{b}}\textrm{b}\overline{\textrm{b}}$ final state using large-area jets in proton-proton collisions at $\sqrt{s}$ = 13 TeV

A search is presented for the resonant production of a pair of standard model-like Higgs bosons using data from proton-proton collisions at a centre-of-mass energy of 13 TeV, collected by the CMS experiment at the CERN LHC in 2016–2018, corresponding to an integrated luminosity of 138 fb −1 . The final state consists of two b quark-antiquark pairs. The search is conducted in the region of phase space where at least one of the pairs is highly Lorentz-boosted and is reconstructed as a single large-area jet. The other pair may be either similarly merged or resolved, the latter reconstructed using two b-tagged jets. The data are found to be consistent with standard model processes and are interpreted as 95% confidence level upper limits on the product of the cross sections and the branching fractions of the spin-0 radion and the spin-2 bulk graviton that arise in warped extradimensional models. The limits set are in the range 9.74–0.29 fb and 4.94–0.19 fb for a narrow radion and a graviton, respectively, with masses between 1 and 3 TeV. For a radion and for a bulk graviton with widths 10% of their masses, the limits are in the range 12.5–0.35 fb and 8.23–0.23 fb, respectively, for the same masses. These limits result in the exclusion of a narrow-width graviton with a mass below 1.2 TeV, and of narrow and 10%-width radions with masses below 2.6, and 2.9 TeV, respectively.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

White Paper: Scalable Digital Twin Capabilities for Aging and Surveillance of Engineered Systems

This white paper presents a multi-year initiative to develop practical, secure, and scalable digital twin capabilities for engineered systems in aging and surveillance contexts—an approach pioneered at the National Nuclear Security Administration (NNSA) Lawrence Livermore National Laboratory (LLNL) that maps directly onto the needs and ambitions of the Navy for ship- and fleet-level digital twins. LLNL’s work in building part- and process-level digital twins for advanced manufacturing, with a vision to scale up to entire factory floors and, ultimately, enterprise-wide digital twins, offers an adaptable pathway for the Navy as it seeks to modernize lifecycle management, readiness, and predictive maintenance across ships and fleets. For our application, we integrate physics-based modeling with automated data ingestion, processing, and AI-driven calibration, creating hybrid models that are both interpretable and data responsive. We modernized legacy workflows, established centralized data infrastructure, automated experimental pipelines, and demonstrated end-to-end coupling of accelerated aging data with finite element simulations via optimization and surrogate modeling. The result is a generalizable framework that supports part-level digital twins today and lays the groundwork for future system-level twins suitable for Navy applications.

36 MATERIALS SCIENCE↗

TPSAS-NF1676L-30074-DND

NASA is at forefront in developing unique active/passive remote sensing capabilities towards space-based observations for understanding the complexities and interactions among Earth system components. The world is facing significant environmental challenges and a robust, integrated, and flexible system of observations and models are needed for understanding the short-and long-term impact on the Earth system. A fundamental challenge for the coming decade is to ensure that space-based observations, analyses, better interpretive understanding, enhanced predictive models, broadened international community participation, and improved means for information assimilation and disseminations are well coordinated to realize the full economic, societal, and security benefit of Earth science. This presentation will provide an overview of active/passive remote sensing technologies and techniques from ground and space towards NASA’s future vision for Earth science missions for global observations, and the challenges associated in applying them for societal benefit.

Upendra N Singh↗

Theoretical interpretation of the HEAO-3 observations of Cygnus X-3 under the HEAO-3 Guest Investigator Program

A model of the galactic X-ray source Cygnus X-3 (Cyg X-3) is presented which deviates from previous models by positing that the X-rays originate in a jet rather than a binary system consiting of an ordinary star and a collapsed object. In the new model, the 4.8 hour period of Cyg X-3 is caused by variable absorption which occurs as the jet precesses. The primary role of the accretion disk corona (ADC) in modulating Cyg X-3 radiation is to make the observed intensity of a blob of material in a jet appear dimmer by absorption. The needed derivation of the positional dependence of the density of the ADC is freed of some complications by assuming that only the inner regions of the disk are precessing, with a period shorter than 4.8 hours. This model permits the secondary star to be a supergiant, as indicated by the luminosity of the system. The model is especially helpful in interpreting production of radio outbursts and very high energy gamma rays.

Marscher, Alan P.↗