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 1,279 records · Page 71

Sensitivities of subgrid-scale physics schemes, meteorological forcing, and topographic radiation in atmosphere-through-bedrock integrated process models: a case study in the Upper Colorado River basin

Abstract. Mountain hydrology is controlled by interacting processes extending from the atmosphere through the bedrock. Integrated process models (IPMs), one of the main tools needed to interpret observations and refine conceptual models of the mountainous water cycle, require meteorological forcing that simulates the atmospheric process to predict hydroclimate then subsequently impacts surface–subsurface hydrology. Complex terrain and extreme spatial heterogeneity in mountainous environments drive uncertainty in several key considerations in IPM configurations and require further quantification and sensitivity analyses. Here, we present an IPM using the Weather Research and Forecasting (WRF) model which forces an integrated hydrologic model, ParFlow-CLM, implemented over a domain centered over the East River watershed (ERW), located in the Upper Colorado River basin (UCRB). The ERW is a heavily instrumented 300 km2 region in the headwaters of the UCRB near Crested Butte, CO, with a growing atmosphere-through-bedrock observation network. Through a series of experiments in the water year 2019 (WY19), we use four meteorological forcings derived from commonly used reanalysis datasets, three subgrid-scale physics scheme configurations in WRF, and two terrain shading options within WRF to test the relative importance of these experimental design choices for key hydrometeorological metrics including precipitation and snowpack, as well as evapotranspiration, groundwater storage, and discharge simulated by the ParFlow-CLM. Our hypothesis is that uncertainty from synoptic-scale forcings produces a much larger spread in surface–subsurface hydrologic fields than subgrid-scale physics scheme choice. Results reveal that the WRF subgrid-scale physics configuration leads to larger spatiotemporal variance in simulated hydrometeorological conditions, whereas variance across meteorological forcing with common subgrid-scale physics configurations is more spatiotemporally constrained. Despite reasonably simulating precipitation, a delay in simulated discharge peak is due to a systematic cold bias across WRF simulations, suggesting the need for bias correction. Discharge shows greater variance in response to the WRF simulations across subgrid-scale physics schemes (26 %) rather than meteorological forcing (6 %). The topographic radiation option has minor effects on the watershed-average hydrometeorological processes but adds profound spatial heterogeneity to local energy budgets (±30 W m−2 in shortwave radiation and 1 K air temperature differences in late summer). This is the first presentation of sensitivity analyses that provide support to help guide the scientific community to develop observational constraints on atmosphere-through-bedrock processes and their interactions.

54 ENVIRONMENTAL SCIENCES↗

Modeling Exposure to Persistent Chemicals in Hazard and Risk Assessment

Fate and exposure modeling has not thus far been explicitly used in the risk profile documents prepared to evaluate significant adverse effect of candidate chemicals for either the Stockholm Convention or the Convention on Long-Range Transboundary Air Pollution. However, we believe models have considerable potential to improve the risk profiles. Fate and exposure models are already used routinely in other similar regulatory applications to inform decisions, and they have been instrumental in building our current understanding of the fate of POP and PBT chemicals in the environment. The goal of this paper is to motivate the use of fate and exposure models in preparing risk profiles in the POP assessment procedure by providing strategies for incorporating and using models. The ways that fate and exposure models can be used to improve and inform the development of risk profiles include: (1) Benchmarking the ratio of exposure and emissions of candidate chemicals to the same ratio for known POPs, thereby opening the possibility of combining this ratio with the relative emissions and relative toxicity to arrive at a measure of relative risk. (2) Directly estimating the exposure of the environment, biota and humans to provide information to complement measurements, or where measurements are not available or are limited. (3) To identify the key processes and chemical and/or environmental parameters that determine the exposure; thereby allowing the effective prioritization of research or measurements to improve the risk profile. (4) Predicting future time trends including how quickly exposure levels in remote areas would respond to reductions in emissions. Currently there is no standardized consensus model for use in the risk profile context. Therefore, to choose the appropriate model the risk profile developer must evaluate how appropriate an existing model is for a specific setting and whether the assumptions and input data are relevant in the context of the application. It is possible to have confidence in the predictions of many of the existing models because of their fundamental physical and chemical mechanistic underpinnings and the extensive work already done to compare model predictions and empirical observations. The working group recommends that modeling tools be applied for benchmarking PBT/POPs according to exposure-to-emissions relationships, and that modeling tools be used to interpret emissions and monitoring data. The further development of models that couple fate, long-range transport, and bioaccumulation should be fostered, especially models that will allow time trends to be scientifically addressed in the risk profile.

54 ENVIRONMENTAL SCIENCES↗

Search for dark matter produced in association with a Higgs boson decaying to tau leptons at $\sqrt{s}$ = 13 TeV with the ATLAS detector

A search for dark matter produced in association with a Higgs boson in final states with two hadronically decaying τ-leptons and missing transverse momentum is presented. The analysis uses 139 fb -1 of proton-proton collision data at $\sqrt{s}$ = 13 TeV collected by the ATLAS experiment at the Large Hadron Collider between 2015 and 2018. No evidence of physics beyond the Standard Model is found. The results are interpreted in terms of a 2HDM+a model featuring two scalar Higgs doublets and a pseudoscalar singlet field. Exclusion limits on the parameters of the model in selected benchmark scenarios are derived at 95% confidence level. Model-independent limits are also set on the visible cross-section for processes beyond the Standard Model producing missing transverse momentum in association with a Higgs boson decaying into τ-leptons.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

A comparison of material flow strength models using Bayesian cross-validation

Predicting material flow strength over a range of conditions, such as temperature and strain rate, is necessary for many engineering applications. In this paper, we consider how to compare the predictiveness of several different strength models using a statistical technique called Bayesian cross-validation. Given a dataset of flow strength measurements obtained from mechanictesting experiments, the procedure consists of performing a Bayesian calibration of each strength model on a subset of the data and evaluating how well the trained models predict the remaining data. The predictiveness of a calibrated strength model is quantifiable probabilistically, which provides an interpretable metric for comparing the different models. As an illustrative example, we compare the Johnson-Cook (JC), Zerilli-Armstrong (ZA), Preston-Tonks-Wallace (PTW), and Mechanical Threshold Stress (MTS) flow strength models for the tantalum stress-strain curve data from Chen and Gray (1996). We show that prediction intervals for the four strength models cover the held-out data at most experimental conditions, but also that prediction interval coverage and prediction uncertainty varies by model and experimental condition. Finally, the analysis further allows us to identify experimental regimes for which one of the strength models predicts better than the other three.

36 MATERIALS SCIENCE↗

Magnetotelluric Investigations of the Kilauea Volcano, Hawaii

In 2002 and 2003 a collaborative effort was undertaken between Lawrence Berkeley National Laboratory, Sandia National Laboratories, the U.S. Geological Survey (USGS) Menlo Park, the USGS Hawaiian Volcano Observatory, and Electromagnetic Instruments Inc. to study the Kilauea volcano in Hawaii using the magnetotelluric (MT) technique. The work was motivated by a desire to improve understanding of the magma reservoirs and conduits within Kilauea and the East and Southwest Rift zones, which has implications for understanding Kilauea's plumbing system. An improved understanding of the rift zones has implications in understanding large-scale landslides that are generated in the Hilina Slump, which produce significant impacts on coastal communities. Up to eight stations operated simultaneously, with multiple remote reference sites, and data were processed using multi-station robust processing techniques. Data was acquired at 70 sites over the Southwest and East rift zones. Good to excellent quality data were obtained even in the harshest conditions, such as those encountered on the fresh lava flows of the East Rift Zone, where electrical contact resistances are on the order of 100 kΩ. A three-dimensional (3D) MT model study was done to guide interpretation of the observed MT measurements. Synthetic modeling demonstrates that conductive bodies in the upper 3 km can be spatially resolved where MT station sampling is good. Resistivity anomalies in the 3D inversions have a high degree of spatial correlation with previously published seismic velocity anomalies beneath Kilauea. Melt fractions between 0.096 and 0.117 are calculated for the Kilauea and Pu'u'ō'ō low resistivity anomalies, respectively.

58 GEOSCIENCES↗

A novel method may reveal bulk metallic glass compressive ductility trends in high data rate nanoindentation

Recent methods allow novel amorphous alloy compositions to be rapidly manufactured at small scale; however, obtaining materials properties such as compressive ductility from these smaller specimens has remained a challenge. Here, we suggest a potential high-throughput nanoindentation method that may be able to rapidly characterize the relative compressive ductility between these alloys based on their serration characteristics. The properties of emergent serrations, when interpreted in a simple micromechanical stress relaxation model, may order these materials by their compressive plastic strain to failure. These results are consistent with the ordering obtained from compressed specimens as well as with model simulations, suggesting that this model may be broadly useful for interpreting compressive ductility from nanoindentation serrations. After it is validated on more materials, this new method will match the rapid pace of amorphous alloy development, thus allowing metallic glass properties to be fine-tuned for each application prior to scale prototyping.

36 MATERIALS SCIENCE↗

Changes in extremely hot days under stabilized 1.5 and 2.0 °C global warming scenarios as simulated by the HAPPI multi-model ensemble

Abstract. The half a degree additional warming, prognosis and projected impacts (HAPPI) experimental protocol provides a multi-model database to compare the effects of stabilizing anthropogenic global warming of 1.5 °C over preindustrial levels to 2.0 °C over these levels. The HAPPI experiment is based upon large ensembles of global atmospheric models forced by sea surface temperature and sea ice concentrations plausible for these stabilization levels. This paper examines changes in extremes of high temperatures averaged over three consecutive days. Changes in this measure of extreme temperature are also compared to changes in hot season temperatures. We find that over land this measure of extreme high temperature increases from about 0.5 to 1.5 °C over present-day values in the 1.5 °C stabilization scenario, depending on location and model. We further find an additional 0.25 to 1.0 °C increase in extreme high temperatures over land in the 2.0 °C stabilization scenario. Results from the HAPPI models are consistent with similar results from the one available fully coupled climate model. However, a complicating factor in interpreting extreme temperature changes across the HAPPI models is their diversity of aerosol forcing changes.

54 ENVIRONMENTAL SCIENCES↗

Model-independent search for pair production of new bosons decaying into muons in proton-proton collisions at $\sqrt{s}$ = 13 TeV

The results of a model-independent search for the pair production of new bosons within a mass range of 0.21 < m < 60 GeV, are presented. This study utilizes events with a four-muon final state. We use two data sets, comprising 41.5 fb −1 and 59.7 fb −1 of proton-proton collisions at $\sqrt{s}$ = 13 TeV, recorded in 2017 and 2018 by the CMS experiment at the CERN LHC. The study of the 2018 data set includes a search for displaced signatures of a new boson within the proper decay length range of 0 < cτ < 100 mm. Our results are combined with a previous CMS result, based on 35.9 fb −1 of proton-proton collisions at $\sqrt{s}$ = 13 TeV collected in 2016. No significant deviation from the expected background is observed. Results are presented in terms of a model-independent upper limit on the product of cross section, branching fraction, and acceptance. The findings are interpreted across various benchmark models, such as an axion-like particle model, a vector portal model, the next-to-minimal supersymmetric standard model, and a dark supersymmetric scenario, including those predicting a non-negligible proper decay length of the new boson. In all considered scenarios, substantial portions of the parameter space are excluded, expanding upon prior results.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

THE CHEMICAL SIGNATURES OF THE FIRST STAR CLUSTERS IN THE UNIVERSE

The chemical abundance patterns of the oldest stars in the Galaxy are expected to contain residual signatures of the first stars in the early universe. Numerous studies attempt to explain the intrinsic abundance scatter observed in some metal-poor populations in terms of chemical inhomogeneities dispersed throughout the early Galactic medium due to discrete enrichment events. Just how the complex data and models are to be interpreted with respect to 'progenitor yields' remains an open question. Here we show that stochastic chemical evolution models to date have overlooked a crucial fact. Essentially, all stars today are born in highly homogeneous star clusters and it is likely that this was also true at early times. When this ingredient is included, the overall scatter in the abundance plane [Fe/H] versus [X/Fe] (C-space), where X is a nucleosynthetic element, can be much less than derived from earlier models. Moreover, for moderately flat cluster mass functions ({gamma} {approx}< 2), and/or for mass functions with a high mass cutoff (M{sub max} {approx}> 10{sup 5} M{sub sun}), stars exhibit a high degree of clumping in C-space that can be identified even in relatively small data samples. Since stellar abundances can be modified by mass transfer in close binaries, clustered signatures are essential for deriving the yields of the first supernovae. We present a statistical test to determine whether a given set of observations exhibit such behavior. Our initial work focuses on two dimensions in C-space, but we show that the clustering signal can be greatly enhanced by additional abundance axes. The proposed experiment will be challenging on existing 8-10 m telescopes, but relatively straightforward for a multi-object echelle spectrograph mounted on a 25-40 m telescope.

79 ASTRONOMY AND ASTROPHYSICS↗

Diagnostics of the solar corona from comparison between Faraday rotation measurements and magnetohydrodynamic simulations

Polarized natural radio sources passing behind the Sun experience Faraday rotation as a consequence of the electron density and magnetic field strength in coronal plasma. Since Faraday rotation is proportional to the product of the density and the component of the magnetic field along the line of sight of the observer, a model is required to interpret the observations and infer coronal structures. Faraday rotation observations have been compared with relatively ad hoc models of the corona. Here for the first time we compare these observations with magnetohydrodynamic (MHD) models of the solar corona driven by measurements of the photospheric magnetic field. We use observations made with the NRAO Very Large Array of 34 polarized radio sources occulted by the solar corona between 5 and 14 solar radii. The measurements were made during 1997 May, and 2005 March and April. We compare the observed Faraday rotation values with values extracted from MHD steady-state simulations of the solar corona. We find that (1) using a synoptic map of the solar magnetic field just one Carrington rotation off produces poorer agreements, meaning that the outer corona changes in the course of one month, even in solar minimum; (2) global MHD models of the solar corona driven by photospheric magnetic field measurements are generally able to reproduce Faraday rotation observations; and (3) some sources show significant disagreement between the model and the observations, which appears to be a function of the proximity of the line of sight to the large-scale heliospheric current sheet.

79 ASTRONOMY AND ASTROPHYSICS↗

Friction- and mountain-torque estimates from global atmospheric data

Seasonal, zonal surface torques between the atmosphere and the earth are estimated and compared, using data from a number of independent sources. The mountain torque is computed both from surface pressure data and from isobaric height data. The friction torque is estimated from the oceanic stress data of Hellerman and Rosenstein. Results for the total torque are inferred from atmospheric angular momentum data. Finally, the globally integrated total torque is compared with astronomical observations of the earth's rotation rate. These comparisons help us to assess the quality of the different results. Zonal torques are also computed using results from a GFDL general circulation model of the atmosphere. A comparison with the corresponding results inferred from real data is presented and interpreted in terms of model accuracy.

Wahr, J. M.↗

F-106 data summary and model results relative to threat criteria and protection design analysis

The NASA F-106 has acquired considerable data on the rates-of-change of electromagnetic parameters on the aircraft surface during 690 direct lightning strikes while penetrating thunderstorms at altitudes ranging from 15,000 to 40,000 feet. These in-situ measurements have provided the basis for the first statistical quantification of the lightning electromagnetic threat to aircrat appropriate for determining lightning indirect effects on aircraft. The data are presently being used in updating previous lightning criteria and standards developed over the years from ground-based measurements. The new lightning standards will, therefore, be the first which reflect actual aircraft responses measured at flight altitudes. The modeling technique developed to interpret and understand the direct strike electromagnetic data acquired on the F-106 provides a means to model the interaction of the lightning channel with the F-106. The reasonable results obtained with the model, compared to measured responses, yield confidence that the model may be credibly applied to other aircraft types and uses in the prediction of internal coupling effects in the design of lightning protection for new aircraft.

Pitts, F. L.↗

Accelerating the design of lattice structures using machine learning

Lattices remain an attractive class of structures due to their design versatility; however, rapidly designing lattice structures with tailored or optimal mechanical properties remains a significant challenge. With each added design variable, the design space quickly becomes intractable. To address this challenge, research efforts have sought to combine computational approaches with machine learning (ML)-based approaches to reduce the computational cost of the design process and accelerate mechanical design. While these efforts have made substantial progress, significant challenges remain in (1) building and interpreting the ML-based surrogate models and (2) iteratively and efficiently curating training datasets for optimization tasks. Here, we address the first challenge by combining ML-based surrogate modeling and Shapley additive explanation (SHAP) analysis to interpret the impact of each design variable. We find that our ML-based surrogate models achieve excellent prediction capabilities (R 2 > 0.95) and SHAP values aid in uncovering design variables influencing performance. We address the second challenge by utilizing active learning-based methods, such as Bayesian optimization, to explore the design space and report a 5 × reduction in simulations relative to grid-based search. Collectively, these results underscore the value of building intelligent design systems that leverage ML-based methods for uncovering key design variables and accelerating design.

36 MATERIALS SCIENCE↗

Mechanical specific energy versus depth of cut in rock cutting and drilling

The relationship between Mechanical Specific Energy (MSE) and the Rate of Penetration (ROP), or equivalently the depth of cut per revolution, provides an important measure for strategizing a drilling operation. This study explores how MSE evolves with depth of cut, and presents a concerted effort that encompasses analytical, computational and experimental approaches. A simple model for the relationship between MSE and cutting depth is first derived with consideration of the wear progression of a circular cutter. This is an extension of Detournay and Defourny's phenomenological cutting model. Wear is modeled as a flat contact area at the bottom of a cutter referred to as a wear flat, and that wear flat in the past is often considered to be fixed during cutting. But during a drilling operation by a full bit that consists of multiple circular cutters, the wear flat length may increase because of various wear mechanisms involved. The wear progression of cutters generally results in reduced efficiency with either increased MSE or decreased ROP. Also, an accurate estimate of removed rock volume is found important for the evaluation of MSE. The derived model is compared with experiment results from a single circular cutter, for cutting soft rock under ambient pressure with actual depth measured through a micrometer, and for cutting high strength rock under high pressure with actual cutting area measured by a confocal microscope. Lastly, the model is employed to interpret the evolution of MSE with depth of cut for a full drilling bit under confining pressure. The general form of equation of the developed model is found to describe well the experiment data and can be applied to interpret the drilling data for a full bit.

42 ENGINEERING↗

Unrolled Video Super-Resolution Network with Autoregressive Prior for the Case of Known Motion

Real-time detection and classification of distant objects is necessary for many national security applications. However, when objects are far from the sensor, they occupy only a small number of pixels in the captured video, limiting the amount of visual detail available for recognition. State-of-the-art classification methods typically rely on high-resolution (HR) video streams to capture characteristic object features, but obtaining such detail is challenging for distant objects that occupy only a few pixels. This motivates the development of video super-resolution (VSR) methods that enhance object classification by recovering fine details from low-pixel representations. Current VSR methods rely either on model-based optimization, which is interpretable but computationally expensive, or on learning-based approaches, which are efficient and high-performing but often lack flexibility and interpretability. In this report, we propose an end-to-end trainable unrolled VSR network, UVSRNet, which super-resolves each frame in a video by exploiting sub-pixel motion between neighboring low-resolution (LR) frames as well as incorporating high-frequency detail from previously super-resolved frames. In particular, by unrolling a plug-and-play (PnP) half-quadratic splitting (HQS) algorithm, we leverage a model-based data-fitting module alongside a learning-based autoregressive prior module. This combination yields a method that maintains the flexibility and interpretability of model-based methods while achieving the performance advantages of learning-based methods.

45 MILITARY TECHNOLOGY, WEAPONRY, AND NATIONAL DEF↗

Treatment with integrase inhibitor suggests a new interpretation of HIV RNA decay curves that reveals a subset of cells with slow integration

The kinetics of HIV-1 decay under treatment depends on the class of antiretrovirals used. Mathematical models are useful to interpret the different profiles, providing quantitative information about the kinetics of virus replication and the cell populations contributing to viral decay. We modeled proviral integration in short- and long-lived infected cells to compare viral kinetics under treatment with and without the integrase inhibitor raltegravir (RAL). We fitted the model to data obtained from participants treated with RAL-containing regimes or with a four-drug regimen of protease and reverse transcriptase inhibitors. Our model explains the existence and quantifies the three phases of HIV-1 RNA decay in RAL-based regimens vs. the two phases observed in therapies without RAL. Our findings indicate that HIV-1 infection is mostly sustained by short-lived infected cells with fast integration and a short viral production period, and by long-lived infected cells with slow integration but an equally short viral production period. We propose that these cells represent activated and resting infected CD4+ T-cells, respectively, and estimate that infection of resting cells represent ~4% of productive reverse transcription events in chronic infection. RAL reveals the kinetics of proviral integration, showing that in short-lived cells the pre-integration population has a half-life of ~7 hours, whereas in long-lived cells this half-life is ~6 weeks. We also show that the efficacy of RAL can be estimated by the difference in viral load at the start of the second phase in protocols with and without RAL. Altogether, we provide a mechanistic model of viral infection that parsimoniously explains the kinetics of viral load decline under multiple classes of antiretrovirals.

59 BASIC BIOLOGICAL SCIENCES↗

Improving parameter estimation for column experiments by multi-model evaluation and comparison

Reliability of column experiment data interpretation by fitting the data to a transport model using the nonlinear least squares method is of concern when the complexity of the experiment increases. Misfits often occur and are accepted with the excuse of system complexity. More often, overfits are published and accepted uncritically. The objective of this work was to improve the reliability of column experiment data interpretation to provide insights for future investigation and select estimates for application prediction. We proposed to achieve this goal by evaluating and comparing various estimates that consider the uncertainties in experimental conditions, sensitivities of model parameters, alternative parameterizations and models. Two examples from the literature were selected for the illustration. Sensitivity analysis showed that the experimentally-controlled parameters (water content, flow rate and pulse volume) are much more sensitive than the transport parameters (dispersivity and mass transfer coefficient) in breakthrough curve fitting. Fixing these sensitive parameters may lead to misfit or biased estimate. Estimating sensitive parameters helps diagnose and remediate the misfit. The dispersion coefficient was parameterized as a function of solute molecular diffusion coefficient, medium dispersivity and average pore water velocity to reduce the number of estimated parameters for multiple tracers. Nonlinear least squares statistics were used to evaluate the goodness-of-fit for the multiple estimates obtained using different fitting schemes (e.g., fixing or estimating parameters), parameterizations and models (e.g., equilibrium or nonequilibrium). The F-test and Akaike s Information Criteria were used for the discrimination of statistically reasonable estimates. Our results suggest that sensitivity/uncertainty analysis and model evaluation /discrimination are useful to improve the reliability of column experiment data interpretation.

58 GEOSCIENCES↗

Combined search for electroweak production of winos, binos, higgsinos, and sleptons in proton-proton collisions at s = 13 TeV

A combination of the results of several searches for the electroweak production of the supersymmetric partners of standard model bosons, and of charged leptons, is presented. All searches use proton-proton collision data at s = 13 TeV recorded with the CMS detector at the LHC in 2016–2018. The analyzed data correspond to an integrated luminosity of up to 137 fb − 1 . The results are interpreted in terms of simplified models of supersymmetry. Two new interpretations are added with this combination: a model spectrum with the bino as the lightest supersymmetric particle together with mass-degenerate Higgsinos decaying to the bino and a standard model boson, and the compressed-spectrum region of a previously studied model of slepton pair production. Improved analysis techniques are employed to optimize sensitivity for the compressed spectra in the wino and slepton pair production models. The results are consistent with expectations from the standard model. The combination provides a more comprehensive coverage of the model parameter space than the individual searches, extending the exclusion by up to 125 GeV, and also targets some of the intermediate gaps in the mass coverage. © 2024 CERN, for the CMS Collaboration 2024 CERN

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗