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 469 records · Page 26

Dynamic models of flux tubes in the interpretation of polarization measurements

Recent observations of Stokes parameter profiles indicate the presence of mass motions with large velocity gradients associated with small-scale magnetic elements. Dynamic models of flux tubes were used in order to interpret observations of unresolved elements. It is clear that the physical picture of the dynamic models will be quite different from the hydrostatic ones since there is a strong coupling between the magnetic and the velocity field. Polarization measurements have to be interpreted in terms of dynamic models. Two-D steady flow solutions in slender magnetic tubes have been worked out. It was found that the main properties of the intensity line profiles as well as the asymmetries of the V Stokes profiles can be explained best in terms of magnetic elements with moderate field strength.

Ribes-Nesme, E.↗

Experimental study of the thermal-acoustic efficiency in a long turbulent diffusion-flame burner

A two-year study of noise production in a long tubular burner is described. The research was motivated by an interest in understanding and eventually reducing core noise in gas turbine engines. The general approach is to employ an acoustic source/propagation model to interpret the sound pressure spectrum in the acoustic far field of the burner in terms of the source spectrum that must have produced it. In the model the sources are assumed to be due uniquely to the unsteady component of combustion heat release; thus only direct combustion-noise is considered. The source spectrum is then the variation with frequency of the thermal-acoustic efficiency, defined as the fraction of combustion heat release which is converted into acoustic energy at a given frequency. The thrust of the research was to study the variation of the source spectrum with the design and operating parameters of the burner.

Mahan, J. R.↗

A comparative study of multimodal data fusion strategies for planetary spectroscopy

Integrating heterogeneous data sources can improve scientific inference when different modalities capture complementary information, but doing so is challenging in high-dimensional, small-sample settings. In spectroscopy for planetary exploration, Laser-Induced Breakdown Spectroscopy (LIBS), Raman Spectroscopy (Raman), Visible Infrared Spectroscopy (VISIR), and Mid-Infrared Spectroscopy (MIR) each examine different aspects of composition and mineralogy, raising fundamental questions about when and how data fusion improves predictive performance. Using a Mars-relevant set of geologic standards with measurements from all four modalities, we present a rigorous systematic evaluation of four data fusion strategies: low-level (data) fusion, mid-level (feature) fusion, high-level (decision) fusion, and residual-boosting (sequential) fusion. We assess performance in predicting oxide composition via nested cross-validation and corrected significance testing to evaluate whether data fusion improves upon single-modality baselines. We show that data fusion does not uniformly improve accuracy, and that observed gains are modest, oxide-dependent, and sensitive to modality and model structure. To move beyond aggregate accuracy metrics, we use model coefficients, permutation importance, and residual gain analysis to examine how the fusion models weight individual modalities and to identify patterns of apparent complementarity or redundancy. Though focused on spectroscopy for planetary exploration, our framework for data fusion evaluation and interpretation extends to other scientific domains with heterogeneous and scarce data and provides a principled approach evaluating data fusion strategies, interpreting modality contributions, and understanding tradeoffs among data fusion strategies.

97 MATHEMATICS AND COMPUTING↗

Systematic study of the uncertainties in fitting the cosmic positron data by AMS-02

The operation of AMS-02 opens a new era for the study of cosmic ray physics with unprecedentedly precise data which are comparable with the laboratory measurements. The high precision data allow a quantitative study on the cosmic ray physics and give strict constraints on the nature of cosmic ray sources. However, the intrinsic errors from the theoretical models to interpret the data become dominant over the errors in the data. In the present work we try to give a systematic study on the uncertainties of the models to explain the AMS-02 positron fraction data, which shows the cosmic ray e{sup +}e{sup −} excesses together with the PAMELA and Fermi-LAT measurements. The excesses can be attributed to contributions from the extra e{sup +}e{sup −} sources, such as pulsars or the dark matter annihilation. The possible systematic uncertainties of the theoretical models considered include the cosmic ray propagation, the treatment of the low energy data, the solar modulation, the pp interaction models, the nuclei injection spectrum and so on. We find that in general a spectral hardening of the primary electron injection spectrum above ∼50–100 GeV is favored by the data. Furthermore, the present model uncertainties may lead to a factor of ∼2 enlargement in the determination of the parameter regions of the extra source, such as the dark matter mass, annihilation rate and so on.

79 ASTRONOMY AND ASTROPHYSICS↗

Development, characterization, and modeling of a high-performance Ru/B2CA catalyst for ammonia synthesis

This paper documents the development and performance of a nano-phase Ru catalyst on a (BaO) x (CaO) y (Al 2 O 3 ) support. Extensive screening of the support’s ternary composition shows the best stoichiometry is (BaO) 2 (CaO)(Al 2 O 3 ), denoted B2CA. The paper first describes catalyst preparation and characterization. The paper reports a detailed 12-step reaction mechanism that represents ammonia synthesis over wide ranges of temperature, pressure, space velocity, and feed composition. Additionally, the mechanism is developed and validated using results of packed-bed experiments. The elementary reaction pathways consider surface adsorbates, including catalyst-poisoning behaviors. The rate expressions include important coverage-dependent activation barriers. Machine learning models assist interpretation of the catalyst-support interactions. The detailed chemistry is much more predictive than is possible with global representations (N 2 + 3H 2 ⇌ 2NH 3 ). The validated models can be applied to assist optimizing reactor design and operating conditions.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Digitalization of an experimental electrochemical reactor via the smart manufacturing innovation platform

The exponential increase in data produced over the last two decades has revolutionized the way we collect, store, process, analyze, model, and interpret information to improve profitability. Manufacturing is no exception. How- ever, Smart Manufacturing, the digital practice, organization, workforce, and infrastructure transformation for collection and deployment of data and models at scale and at all levels of manufacturing, is a complex, costly, and labor-intensive journey that is still seeing slow adoption. The Clean Energy Smart Manufacturing Innovation Institute (CESMII), a national Manufacturing USA public-private partnership sponsored by the Department of Energy, is addressing this scaled use of data and modeling in manufacturing. CESMII has focused on how to col- lect and use operating data for numerous applications that improve productivity, precision, and performance of manufacturing operations from factory floor to supply chain using process simulation, predictive analytics, mon- itoring and control, and real-time optimization. Because contextualized data are key, CESMII has developed the Smart Manufacturing Innovation Platform (SMIP) to lower the barriers to the data that are needed to accelerate data-based model building, improve data visualization, and more quickly gain insights. Reusable, standards-based ways of doing data collection, ingestion, and contextualization are particularly important for scaling access and use of data. The SMIP uses a standards-based definition and construct for reusable information models called an SM Profile. When an SM Profile is used in conjunction with the SMIP, the SMIP ensures the availability of contextualized, operational data for model building. The present work demonstrates Smart Manufacturing and the application of the SMIP for building several data-centered models for the operation and control of an ex- perimental electrochemical reactor that reduces carbon dioxide (CO 2 ) gas to valuable liquid and gas chemicals, such as alcohols, olefins, and syngas. We describe how the SMIP plays a central role in more effective model building and we demonstrate how the electochemical reactor can be controlled and optimized for the desired products. Use of the SMIP involves the transmission of real-time sensor measurements to a cloud resource so that the operating data are available to all model building experts. The data collection and transmission process is fully automated to greatly reduce the need for manual manipulation of the data. Data-driven machine learning models are used for advanced real-time state estimation, real-time optimization, and model-based feedback control for the reactor. The application models are implemented as a system to monitor the data flow and control the electrochemical reactor with a single visualization interface. SM Profiles are used to demonstrate reusability of the information models for the reactor and the instrumentation. The application packages, algorithms, and user interfaces developed are cast as Docker images in a library to facilitate reusability of the application models.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Revising the dynamic energy budget theory with a new reserve mobilization rule and three example applications to bacterial growth

Dynamic energy budget (DEB) theory has been applied to model a wide range of organisms, including microbes. In the standard DEB model, biomass is partitioned into reserve and structural compartments, where reserve biomass is mobilized in a pseudolinear manner (while the reserve biomass density, defined as the ratio between reserve and structural biomass, decays linearly) to drive maintenance and the growth of structural biomass (and extracellular enzyme production if it is considered). However, the linear dynamics of the reserve biomass density makes the standard DEB model incapable of explaining the slowdown of microbial growth at high reserve density that is caused by macromolecular crowding effect which reduces biochemical reaction rates (a typical situation occurs when microbes are experiencing severe moisture stress) and is inconsistent with the observation that intracellular enzymatic reactions generally follow non-linear kinetics. By partitioning biomass into reserve, kinetic, and structural compartments, we show here that the Equilibrium Chemistry Approximation (ECA) kinetics can be used to represent enzymatically catalyzed reserve biomass mobilization that can then drive the kinetic and structural biomass synthesis. This revised DEB model better represents the tradeoff in ribosome allocation for structural growth and internal enzyme production, is structurally compatible with metabolic models of cell individuals, and includes the standard DEB model and the popular compromise model as special cases for representing population growth. We then applied the revised DEB model to interpret components of bacterial respiration, their dependence on substrate availability, and emergent microbial carbon use efficiency dynamics for an exponentially growing population. We found that the revised DEB model enables a better understanding of bacterial substrates use (carbon in our examples) than that can be derived from a few other models in the literature. In particular, the revised DEB model explains why carbon use efficiency may first increase, then plateau, and finally decrease with growth rate (and substrate uptake rate), as a function of proteomics. Additionally, the revised DEB model explains why the kinetic biomass compartment needs to be divided to reasonably incorporate proteomic control of microbial growth.

59 BASIC BIOLOGICAL SCIENCES↗

Experimental evidence for an optical interference model for vibrational sum frequency generation on multilayer organic thin film systems. I. Electric dipole approximation

In the field of vibrational sum frequency generation spectroscopy (VSFG) applied to organic thin film systems, a significant challenge to data analysis is in the accurate description of optical interference effects. Herein, we provide experimental evidence that a model recently developed in our lab provides an accurate description of this phenomenon. We studied the organic small molecule N,N′-dioctyl-3,4,9,10-perylenedicarboximide vapor deposited as a thickness gradient on silicon wafer substrates with two oxide thicknesses and two surface preps. VSFG data were obtained using the ssp and the sps polarization combinations in the imide carbonyl stretching region as a function of organic thickness. In this first of two reports, the data are modeled and interpreted within the ubiquitous electric dipole approximation for VSFG. The intrinsic sample responses are parameterized during the fitting routines while optical interference effects are simply calculated from the model using known refractive indices, thin film thicknesses, and beam angles. The results indicate that the thin film model provides a good description of optical interferences, indicating that interfacial terms are significant. Inconsistencies between the fitting results within the bounds of the electric dipole response motivate deliberation for additional effects to be considered in the second report.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Rethinking Sensitivity Analysis of Nuclear Simulations with Topology

In nuclear engineering, understanding the safety margins of the nuclear reactor via simulations is arguably of paramount importance in predicting and preventing nuclear accidents. It is therefore crucial to perform sensitivity analysis to understand how changes in the model inputs affect the outputs. Modern nuclear simulation tools rely on numerical representations of the sensitivity information -- inherently lacking in visual encodings -- offering limited effectiveness in communicating and exploring the generated data. In this paper, we design a framework for sensitivity analysis and visualization of multidimensional nuclear simulation data using partition-based, topology-inspired regression models and report on its efficacy. We rely on the established Morse-Smale regression technique, which allows us to partition the domain into monotonic regions where easily interpretable linear models can be used to assess the influence of inputs on the output variability. The underlying computation is augmented with an intuitive and interactive visual design to effectively communicate sensitivity information to the nuclear scientists. Our framework is being deployed into the multi-purpose probabilistic risk assessment and uncertainty quantification framework RAVEN (Reactor Analysis and Virtual Control Environment). We evaluate our framework using an simulation dataset studying nuclear fuel performance.

computational topology↗

Di-CNN: Domain-Knowledge-Informed Convolutional Neural Network for Manufacturing Quality Prediction

In manufacturing, convolutional neural networks (CNNs) are widely used on image sensor data for data-driven process monitoring and quality prediction. However, as purely data-driven models, CNNs do not integrate physical measures or practical considerations into the model structure or training procedure. Consequently, CNNs’ prediction accuracy can be limited, and model outputs may be hard to interpret practically. This study aims to leverage manufacturing domain knowledge to improve the accuracy and interpretability of CNNs in quality prediction. A novel CNN model, named Di-CNN, was developed that learns from both design-stage information (such as working condition and operational mode) and real-time sensor data, and adaptively weighs these data sources during model training. It exploits domain knowledge to guide model training, thus improving prediction accuracy and model interpretability. A case study on resistance spot welding, a popular lightweight metal-joining process for automotive manufacturing, compared the performance of (1) a Di-CNN with adaptive weights (the proposed model), (2) a Di-CNN without adaptive weights, and (3) a conventional CNN. The quality prediction results were measured with the mean squared error (MSE) over sixfold cross-validation. Model (1) achieved a mean MSE of 6.8866 and a median MSE of 6.1916, Model (2) achieved 13.6171 and 13.1343, and Model (3) achieved 27.2935 and 25.6117, demonstrating the superior performance of the proposed model.

47 OTHER INSTRUMENTATION↗

KINETIC SIMULATION OF SLOW MAGNETOSONIC WAVES AND QUASI-PERIODIC UPFLOWS IN THE SOLAR CORONA

Quasi-periodic disturbances of emission-line parameters are frequently observed in the corona. These disturbances propagate upward along the magnetic field with speeds of ∼100 km s{sup −1}. This phenomenon has been interpreted as evidence of the propagation of slow magnetosonic waves or has been argued to be a signature of intermittent outflows superposed on the background plasmas. Here we aim to present a new “wave + flow” model to interpret these observations. In our scenario, the oscillatory motion is a slow-mode wave, and the flow is associated with a beam created by the wave–particle interaction owing to Landau resonance. With the help of a kinetic model, we simulate the propagation of slow-mode waves and the generation of beam flows. We find that weak periodic beam flows can be generated by to Landau resonance in the solar corona, and the phase with the strongest blueward asymmetry is ahead of that with the strongest blueshift by about 1/4 period. We also find that the slow wave damps to the level of 1/ e after the transit time of two wave periods, owing to Landau damping and Coulomb collisions in our simulation. This damping timescale is similar to that resulting from thermal conduction in the MHD regime. The beam flow is weakened/attenuated with increasing wave period and decreasing wave amplitude since Coulomb collisions become more and more dominant over the wave action. We suggest that this “wave + flow” kinetic model provides an alternative explanation for the observed quasi-periodic propagating perturbations in various parameters in the solar corona.

79 ASTRONOMY AND ASTROPHYSICS↗

Analysis of EIT/LASCO Observations Using Available MHD Models: Investigation of CME Initiation Propagation and Geoeffectiveness

The Sun's activity drives the variability of geospace (i.e., near-earth environment). Observations show that the ejection of plasma from the sun, called coronal mass ejections (CMEs), are the major cause of geomagnetic storms. This global-scale solar dynamical feature of coronal mass ejection was discovered almost three decades ago by the use of space-borne coronagraphs (OSO-7, Skylab/ATM and P78-1). Significant progress has been made in understanding the physical nature of the CMEs. Observations show that these global-scale CMEs have size in the order of a solar radius (approximately 6.7 x 10(exp 5) km) near the sun, and each event involves a mass of about 10(exp 15) g and an energy comparable to that of a large flare on the order of 10(exp 32) ergs. The radial propagation speeds of CMEs have a wide range from tens to thousands of kilometers per second. Thus, the transit time to near earth's environment [i.e., 1 AU (astronomical unit)] can be as fast as 40 hours to 100 hours. The typical transit time for geoeffective events is approximately 60-80 h. This paper consists of two parts: 1) A summary of the observed CMEs from Skylab to the present SOHO will be presented. Special attention will be made to SOHO/ LASCO/ EIT observations and their characteristics leading to a geoeffectiv a CME 2) The chronological development of theory and models to interpret the physical nature of this fascinating phenomenon will be reviewed. Finally, an example will be presented to illustrate the geoeffectiveness of the CMEs by using both observation and model.

Wu, S. T.↗

SULI Oral Presentation

Furthering our understanding of the prevalence and severity of issues that customers face when charging their electric vehicles (EVs) is crucial in order to improve the charging experience across the United States. This project utilizes web-scraping, machine leaning (ML), and natural language processing (NLP) techniques to analyze and categorize user-generated reviews. Selenium was used to build a data collection tool that can scrape vast amounts of user review data from the PlugShare website. Sentiment analysis was employed on this dataset in order to filter out negative reviews for further analysis. NLP techniques such as tokenization and word embedding were then used to convert user-written comments into a numerical format that a ML model can interpret. Multiple ML approaches are currently being explored in order to identify and categorize the charging issues being talked about in each review. Ultimately, the results from the ML model will be visualized and explained in a report on customer pain points to be delivered to the ChargeX Consortium, therefore revealing specific areas for improvement in the customer charging experience.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

SULI Oral Presentation

Furthering our understanding of the prevalence and severity of issues that customers face when charging their electric vehicles (EVs) is crucial in order to improve the charging experience across the United States. This project utilizes web-scraping, machine leaning (ML), and natural language processing (NLP) techniques to analyze and categorize user-generated reviews. Selenium was used to build a data collection tool that can scrape vast amounts of user review data from the PlugShare website. Sentiment analysis was employed on this dataset in order to filter out negative reviews for further analysis. NLP techniques such as tokenization and word embedding were then used to convert user-written comments into a numerical format that a ML model can interpret. Multiple ML approaches are currently being explored in order to identify and categorize the charging issues being talked about in each review. Ultimately, the results from the ML model will be visualized and explained in a report on customer pain points to be delivered to the ChargeX Consortium, therefore revealing specific areas for improvement in the customer charging experience.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

Incorporating Suborganismal Processes into Dynamic Energy Budget Models for Ecological Risk Assessment

Here, a working group at the National Institute for Mathematical and Biological Synthesis (NIMBioS) explored the feasibility of integrating 2 complementary approaches relevant to ecological risk assessment. Adverse outcome pathway (AOP) models provide “bottom-up” mechanisms to predict specific toxicological effects that could affect an individual's ability to grow, reproduce, and/or survive from a molecular initiating event. Dynamic energy budget (DEB) models offer a “top-down” approach that reverse engineers stressor effects on growth, reproduction, and/or survival into modular characterizations related to the acquisition and processing of energy resources. Thus, AOP models quantify linkages between measurable molecular, cellular, or organ-level events, but they do not offer an explicit route to integratively characterize stressor effects at higher levels of organization. While DEB models provide the inherent basis to link effects on individuals to those at the population and ecosystem levels, their use of abstract variables obscures mechanistic connections to suborganismal biology. To take advantage of both approaches, we developed a conceptual model to link DEB and AOP models by interpreting AOP key events as measures of damage-inducing processes affecting DEB variables and rates. We report on the type and structure of data that are generated for AOP models that may also be useful for DEB models. We also report on case studies under development that merge information collected for AOPs with DEB models and highlight some of the challenges. Finally, we discuss how the linkage of these 2 approaches can improve ecological risk assessment, with possibilities for progress in predicting population responses to toxicant exposures within realistic environments.

54 ENVIRONMENTAL SCIENCES↗

Search for electroweak production of charginos in final states with two τ leptons in pp collisions at $ \sqrt{s}=8 $ TeV

Results are presented from a search for the electroweak production of supersymmetric particles in pp collisions in final states with two τ leptons. The data sample corresponds to an integrated luminosity between 18.1 fb$^{−1}$ and 19.6 fb$^{−1}$ depending on the final state of τ lepton decays, at $ \sqrt{s}=8 $ TeV, collected by the CMS experiment at the LHC. The observed event yields in the signal regions are consistent with the expected standard model backgrounds. The results are interpreted using simplified models describing the pair production and decays of charginos or τ sleptons. For models describing the pair production of the lightest chargino, exclusion regions are obtained in the plane of chargino mass vs. neutralino mass under the following assumptions: the chargino decays into third-generation sleptons, which are taken to be the lightest sleptons, and the sleptons masses lie midway between those of the chargino and the neutralino. Chargino masses below 420 GeV are excluded at a 95% confidence level in the limit of a massless neutralino, and for neutralino masses up to 100 GeV, chargino masses up to 325 GeV are excluded at 95% confidence level. Constraints are also placed on the cross section for pair production of τ sleptons as a function of mass, assuming a massless neutralino.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Search for new physics with the M T2 variable in all-jets final states produced in pp collisions at $ \sqrt{s}=13 $ TeV

A search for new physics is performed using events that contain one or more jets, no isolated leptons, and a large transverse momentum imbalance, as measured through the M T2 variable, which is an extension of the transverse mass in events with two invisible particles. The results are based on a sample of proton-proton collisions collected at a center-of-mass energy of 13 TeV with the CMS detector at the LHC, and that corresponds to an integrated luminosity of 2.3 fb –1 . The observed event yields in the data are consistent with predictions for the standard model backgrounds. The results are interpreted using simplified models of supersymmetry and are expressed in terms of limits on the masses of potential new colored particles. Assuming that the lightest neutralino is stable and has a mass less than about 500 GeV, gluino masses up to 1550-1750 GeV are excluded at 95% confidence level, depending on the gluino decay mechanism. For the scenario of direct production of squark-antisquark pairs, top squarks with masses up to 800 GeV are excluded, assuming a 100% branching fraction for the decay to a top quark and neutralino. Moreover, bottom squark masses are excluded up to 880 GeV, and masses of light-flavor squarks are excluded up to 600-1260 GeV, depending on the degree of degeneracy of the squark masses.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Data-driven high-dimensional statistical inference with generative models

Crucial to many measurements at the LHC is the use of correlated multi-dimensional information to distinguish rare processes from large backgrounds, which is complicated by the poor modeling of many of the crucial backgrounds in Monte Carlo simulations. In this work, we introduce HI-SIGMA, a method to perform unbinned high-dimensional statistical inference with data-driven background distributions. In contradistinction to many applications of Simulation Based Inference in High Energy Physics, HI-SIGMA relies on generative ML models, rather than classifiers, to learn the signal and background distributions in the high-dimensional space. These ML models allow for interpretable inference while also incorporating model errors and other sources of systematic uncertainties. We showcase this methodology on a simplified version of a di-Higgs measurement in the bbγγ final state, where the di-photon resonance allows for background interpolation from sidebands into the signal region. We demonstrate that HI-SIGMA provides improved sensitivity as compared to standard classifier-based methods, and that systematic uncertainties can be straightforwardly incorporated by extending methods which have been used for histogram based analyses.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗