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 1,585 records · Page 88

Observation and integrated Earth-system science: A roadmap for 2016–2025

This report is the response to a request by the Committee on Space Research of the International Council for Science to prepare a roadmap on observation and integrated Earth-system science for the coming ten years. Its focus is on the combined use of observations and modelling to address the functioning, predictability and projected evolution of interacting components of the Earth system on timescales out to a century or so. It discusses how observations support integrated Earth-system science and its applications, and identifies planned enhancements to the contributing observing systems and other requirements for observations and their processing. All types of observation are considered, but emphasis is placed on those made from space. The origins and development of the integrated view of the Earth system are outlined, noting the interactions between the main components that lead to requirements for integrated science and modelling, and for the observations that guide and support them. What constitutes an Earth-system model is discussed. Summaries are given of key cycles within the Earth system. The nature of Earth observation and the arrangements for international coordination essential for effective operation of global observing systems are introduced. Instances are given of present types of observation, what is already on the roadmap for 2016–2025 and some of the issues to be faced. Observations that are organized on a systematic basis and observations that are made for process understanding and model development, or other research or demonstration purposes, are covered. Specific accounts are given for many of the variables of the Earth system. The current status and prospects for Earth-system modelling are summarized. The evolution towards applying Earth-system models for environmental monitoring and prediction as well as for climate simulation and projection is outlined. General aspects of the improvement of models, whether through refining the representations of processes that are already incorporated or through adding new processes or components, are discussed. Some important elements of Earth-system models are considered more fully. Data assimilation is discussed not only because it uses observations and models to generate datasets for monitoring the Earth system and for initiating and evaluating predictions, in particular through reanalysis, but also because of the feedback it provides on the quality of both the observations and the models employed. Inverse methods for surface-flux or model-parameter estimation are also covered. Reviews are given of the way observations and the processed datasets based on them are used for evaluating models, and of the combined use of observations and models for monitoring and interpreting the behaviour of the Earth system and for predicting and projecting its future. A set of concluding discussions covers general developmental needs, requirements for continuity of space-based observing systems, further long-term requirements for observations and other data, technological advances and data challenges, and the importance of enhanced international co-operation.

Earth system science↗

Measurement of the top quark mass in the dilepton final state using the matrix element method

The top quark, discovered in 1995 by the CDF and D0 experiments at the Fermilab Tevatron Collider, is the heaviest known fundamental particle. The precise knowledge of its mass yields important constraints on the mass of the yet-unobserved Higgs boson and allows to probe for physics beyond the Standard Model. The first measurement of the top quark mass in the dilepton channel with the Matrix Element method at the D0 experiment is presented. After a short description of the experimental environment and the reconstruction chain from hits in the detector to physical objects, a detailed review of the Matrix Element method is given. The Matrix Element method is based on the likelihood to observe a given event under the assumption of the quantity to be measured, e.g. the mass of the top quark. The method has undergone significant modifications and improvements compared to previous measurements in the lepton+jets channel: the two undetected neutrinos require a new reconstruction scheme for the four-momenta of the final state particles, the small event sample demands the modeling of additional jets in the signal likelihood, and a new likelihood is designed to account for the main source of background containing tauonic Z decay. The Matrix Element method is validated on Monte Carlo simulated events at the generator level. For the measurement, calibration curves are derived from events that are run through the full D0 detector simulation. The analysis makes use of the Run II data set recorded between April 2002 and May 2008 corresponding to an integrated luminosity of 2.8 fb -1 . A total of 107 t$\bar{t}$ candidate events with one electron and one muon in the final state are selected. Applying the Matrix Element method to this data set, the top quark mass is measured to be m top Run IIa = 170.6 ± 6.1(stat.) -1.5 +2.1 (syst.)GeV; m top Run IIb = 174.1 ± 4.4(stat.) -1.8 +2.5 (syst.)GeV; m top comb = 172.9 ± 3.6(stat.) ± 2.3(syst.)GeV. Systematic uncertainties are discussed, and the results are interpreted within the Standard Model of particle physics. As the main systematic uncertainty on the top quark mass comes from the knowledge of the absolute jet energy scale, studies for a simultaneous measurement of the top quark mass and the b jet energy scale are presented. The prospects that such a simultaneous determination offer for future measurements of the top quark mass are outlined.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Modeling a High-Temperature Electrochemically Driven Water-Gas-Shift Process Using a Mixed-Conducting Membrane without External Electrical Power

This paper develops a model to predict and interpret the performance of an elevated-temperature, electrochemical, membrane-assisted, water-gas-shift process. The process uses separated feed streams of H 2 O and CO to produce separated streams of H 2 and CO 2 , without an external electrical power source. The dense ceramic membrane is mixed ionic-electronic-conducting (MIEC) gadolinium-doped ceria (GDC) and the porous composite electrodes are Ni-YSZ. At elevated temperature, GDC conducts both oxygen ions and small polarons. The present process uses chemical potential to drive the process. Electrochemical oxidation of CO proceeds within the composite anode and H 2 O reduction proceeds within the composite cathode. At high temperature (e.g., T > 700 °C), GDC has significant electronic leakage in the form of a reduced-cerium small polaron, which supports the charge-transfer reactions. In a typical electrolyzer or fuel cell, this leakage is significantly problematic. However, the present process depends on the leakage current to complete the electrochemical circuit. Model development and validation is based on measured material properties and reactor performance. Potential applications include using CO-rich blast-furnace off gases in steel processing, producing separated streams of H 2 and CO 2 .

Zhu, Huayang↗

Investigation of the Polytropic Relationship Between Density and Temperature Within Interplanetary Coronal Mass Ejections Using Numerical Simulations

Single-point spacecraft measurements within coronal mass ejections (CMEs) often exhibit a negative correlation between electron density and temperature. At least two opposing interpretations have been suggested for this relationship. If, on one hand, these single spacecraft observations provide direct measures of the polytropic properties of the plasma, then they imply that the polytropic index for the electrons gamma(sub e) is often < 1. Moreover, since the electrons carry the bulk of the pressure (via their significantly higher temperature), this further implies that the dynamics of CME evolution are dominated by an effective polytropic index gamma(sub e)ff < 1. On the other hand, gamma < 1 implies that as the ejecta propagate away from the Sun and expand, they also heat up; a result clearly at odds with in situ observations. In contrast to these CME intervals, many studies have shown that the quiescent solar wind exhibits a positive correlation between electron density and temperature, suggesting that gamma(sub e) > 1. In this study we simulate the evolution of a variety of CME-like disturbances in the solar wind using a one-dimensional, single-fluid model, to address the interpretation of the relationship between electron density and temperature within CMEs at fixed locations in space. Although we strictly impose a polytropic relationship (with gamma = constant) throughout our simulations, we demonstrate that a variety of correlations can exist between density and temperature at fixed points. Furthermore, we demonstrate that the presence of only local uncorrelated random fluctuations in density and temperature can produce a negative correlation. Consequently, we conclude that these single-point observations of negative correlations between electron density and temperature cannot be used to infer the value of gamma(sub e). Instead, we suggest that entropy variations, together with the plasma's tendency to achieve pressure balance with its surroundings, are responsible for the observed profiles.

Riley, Pete↗

Deep neural network improves the estimation of polygenic risk scores for breast cancer

Polygenic risk scores (PRS) estimate the genetic risk of an individual for a complex disease based on many genetic variants across the whole genome. Here, we compared a series of computational models for estimation of breast cancer PRS. A deep neural network (DNN) was found to outperform alternative machine learning techniques and established statistical algorithms, including BLUP, BayesA, and LDpred. In the test cohort with 50% prevalence, the Area Under the receiver operating characteristic Curve (AUC) were 67.4% for DNN, 64.2% for BLUP, 64.5% for BayesA, and 62.4% for LDpred. BLUP, BayesA, and LPpred all generated PRS that followed a normal distribution in the case population. However, the PRS generated by DNN in the case population followed a bimodal distribution composed of two normal distributions with distinctly different means. This suggests that DNN was able to separate the case population into a high-genetic-risk case subpopulation with an average PRS significantly higher than the control population and a normal-genetic-risk case subpopulation with an average PRS similar to the control population. This allowed DNN to achieve 18.8% recall at 90% precision in the test cohort with 50% prevalence, which can be extrapolated to 65.4% recall at 20% precision in a general population with 12% prevalence. Interpretation of the DNN model identified salient variants that were assigned insignificant p values by association studies, but were important for DNN prediction. These variants may be associated with the phenotype through nonlinear relationships.

59 BASIC BIOLOGICAL SCIENCES↗

Feedback and sensitivity in an electrical circuit: An analog for climate models

Earth's climate sensitivity is often interpreted in terms of feedbacks that can alter the sensitivity from that of a no-feedback Stefan-Boltzmann radiator, with the feedback concept and algebra introduced by analogy to the use of this concept in the electronics literature. This analogy is quite valuable in interpreting the sensitivity of the climate system, but usage of this algebra and terminology in the climate literature is often inconsistent, with resultant potential for confusion and loss of physical insight. Here a simple and readily understood electrical resistance circuit is examined in terms of feedback theory to introduce and define the terminology that is used to quantify feedbacks. This formalism is applied to the feedbacks in an energy-balance model of Earth's climate and used to interpret the magnitude of feedback in the climate system that corresponds to present estimates of Earth's climate sensitivity.

54 ENVIRONMENTAL SCIENCES↗

Techniques and tactics used in determining the structure of the trimeric ebolavirus glycoprotein

Here, the techniques, tactics and strategies used to overcome a series of technical roadblocks in crystallization and phasing of the trimeric ebolavirus glycoprotein are described. The trimeric membrane-anchored ebolavirus envelope glycoprotein (GP) is responsible for viral attachment, fusion and entry. Knowledge of its structure is important both for understanding ebolavirus entry and for the development of medical interventions. Crystal structures of viral glycoproteins, especially those in their metastable prefusion oligomeric states, can be difficult to achieve given the challenges in production, purification, crystallization and diffraction that are inherent in the heavily glycosylated flexible nature of these types of proteins. The crystal structure of ebolavirus GP in its trimeric prefusion conformation in complex with a human antibody derived from a survivor of the 1995 Kikwit outbreak has now been determined [Lee et al. (2008 ▶), Nature (London), 454, 177–182]. Here, the techniques, tactics and strategies used to overcome a series of technical roadblocks in crystallization and phasing are described. Glycoproteins were produced in human embryonic kidney 293T cells, which allowed rapid screening of constructs and expression of protein in milligram quantities. Complexes of GP with an antibody fragment (Fab) promoted crystallization and a series of deglycosylation strategies, including sugar mutants, enzymatic deglycosylation, insect-cell expression and glycan anabolic pathway inhibitors, were attempted to improve the weakly diffracting glycoprotein crystals. The signal-to-noise ratio of the search model for molecular replacement was improved by determining the structure of the uncomplexed Fab. Phase combination with Fab model phases and a selenium anomalous signal, followed by NCS-averaged density modification, resulted in a clear interpretable electron-density map. Model building was assisted by the use of B-value-sharpened electron-density maps and the proper sequence register was confirmed by building alternate sequences using N-linked glycan sites as anchors and secondary-structural predictions.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Search for supersymmetry in pp collisions at $\sqrt{s} =7$ TeV in events with a single lepton, jets, and missing transverse momentum

Results are reported from a search for new physics processes in events containing a single isolated high-transverse-momentum lepton (electron or muon), energetic jets, and large missing transverse momentum. The analysis is based on a 4.98 fb -1 sample of proton–proton collisions at a center-of-mass energy of 7 TeV, obtained with the CMS detector at the LHC. Three separate background estimation methods, each relying primarily on control samples in the data, are applied to a range of signal regions, providing complementary approaches for estimating the background yields. The observed yields are consistent with the predicted standard model backgrounds. The results are interpreted in terms of limits on the parameter space for the constrained minimal supersymmetric extension of the standard model, as well as on cross sections for simplified models, which provide a generic description of the production and decay of new particles in specific, topology based final states.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Using Ultrasound Image Augmentation and Ensemble Predictions to Prevent Machine-Learning Model Overfitting

Deep learning predictive models have the potential to simplify and automate medical imaging diagnostics by lowering the skill threshold for image interpretation. However, this requires predictive models that are generalized to handle subject variability as seen clinically. Here, we highlight methods to improve test accuracy of an image classifier model for shrapnel identification using tissue phantom image sets. Using a previously developed image classifier neural network—termed ShrapML—blind test accuracy was less than 70% and was variable depending on the training/test data setup, as determined by a leave one subject out (LOSO) holdout methodology. Introduction of affine transformations for image augmentation or MixUp methodologies to generate additional training sets improved model performance and overall accuracy improved to 75%. Further improvements were made by aggregating predictions across five LOSO holdouts. This was done by bagging confidences or predictions from all LOSOs or the top-3 LOSO confidence models for each image prediction. Top-3 LOSO confidence bagging performed best, with test accuracy improved to greater than 85% accuracy for two different blind tissue phantoms. This was confirmed by gradient-weighted class activation mapping to highlight that the image classifier was tracking shrapnel in the image sets. Overall, data augmentation and ensemble prediction approaches were suitable for creating more generalized predictive models for ultrasound image analysis, a critical step for real-time diagnostic deployment.

60 APPLIED LIFE SCIENCES↗

Comparison of Equilibrium Climate Sensitivity Estimates From Slab Ocean, 150‐Year, and Longer Simulations

We compare equilibrium climate sensitivity (ECS) estimates from pairs of long (≥ 800‐year) control and abruptly quadrupled CO2 simulations with shorter (150, 300 year) coupled atmosphere‐ocean simulations and Slab Ocean Models (SOM). Consistent with previous work, ECS estimates from shorter coupled simulations based on annual averages for years 1‐150 underestimate those from SOM (‐8% ± 13%) and long (‐14% ± 8%) simulations. Analysis of only years 21‐150 improved agreement with SOM (‐2% ± 14%) and long (‐8% ± 10%) estimates. Use of pentadal averages for years 51‐150 results in improved agreement with long simulations (‐4% ± 11%). While ECS estimates from current generation US models based on SOM and coupled annual averages of years 1‐150 range from 2.6°C to 5.3°C, estimates based longer simulations of the same models range from 3.2°C to 7.0°C. Such variations between methods argues for caution in comparison and interpretation of ECS estimates across models.

Climate sensitivity↗

MONITORING AND MODELLING OF RADIOLYTIC DEGRADATION PRODUCTS OF TBP/n-DODECANE

The Plutonium Uranium Extraction (PUREX) solvent system was developed for the separation of plutonium and uranium from irradiated fuel. Since the implementation of this process, the degradation chemistry associated with the irradiated solvent system, tributyl phosphate (TBP)/n-dodecane/nitric acid has been extensively studied as the integrity of the organic solvent is paramount for reproducible performance of the separation flowsheet (extraction/scrub/strip) during multiple cycles. In PUREX-like processes, the extent of decomposition is dependent not only upon the solvent, but also upon the presence of constituents, such as nitric acid, that interact with TBP and increase its susceptibility to radiolytic degradation. The build-up of degradation products in the organic phase alters process flowsheet performance via modification of the metal ions speciation, loss of solvent components, and enhanced water transport into the organic phase. On-line identification and quantification of the solvent degradation products would provide the necessary information for more detailed process control as well as providing the basis for timing solvent washing or replacement. In our research, we are exploring the potential of on-line monitoring for the PUREX solvent radiolytic degradation products. To identify degradation products, TBP/n-dodecane solvent, contacted with aqueous nitric acid solutions of variable concentrations are subjected to various gamma radiation external doses then analyzed by electro-spray ionization mass spectrometry (ESMS). In addition, vibrational spectroscopy is utilized to monitor and quantify major degradation products including dibutyl phosphoric acid (HDBP) and monobutyl phosphoric acid (H2MBP) in TBP/n-dodecane solvent. The compiled spectroscopic databases serve for developing interpretive and predictive chemometric models for the quantification of the PUREX solvent degradation products.

Peterson, James M.↗

ACES-GNN: can graph neural network learn to explain activity cliffs?

Graph Neural Networks (GNNs) have revolutionized molecular property prediction by leveraging graph-based representations, yet their opaque decision-making processes hinder broader adoption in drug discovery. This study introduces the Activity-Cliff-Explanation-Supervised GNN (ACES-GNN) framework, designed to simultaneously improve predictive accuracy and interpretability by integrating explanation supervision for activity cliffs (ACs) into GNN training. ACs, defined by structurally similar molecules with significant potency differences, pose challenges for traditional models due to their reliance on shared structural features. By aligning model attributions with chemist-friendly interpretations, the ACES-GNN framework bridges the gap between prediction and explanation. Validated across 30 pharmacological targets, ACES-GNN consistently enhances both predictive accuracy and attribution quality for ACs compared to unsupervised GNNs. Our results demonstrate a positive correlation between improved predictions and accurate explanations, offering a robust and adaptable framework to better understand and interpret ACs. This work underscores the potential of explanation-guided learning to advance interpretable artificial intelligence in molecular modeling and drug discovery.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

High-throughput Li plating quantification for fast-charging battery design

Fast charging of most commercial lithium-ion batteries is limited due to fear of lithium plating on the graphite anode, which is difficult to detect and poses considerable safety risk. In this report we demonstrate the power of simple, accessible and high-throughput cycling techniques to quantify irreversible Li plating spanning data from over 200 cells. We first observe the effects of energy density, charge rate, temperature and state of charge on lithium plating, use the results to refine a mature physics-based electrochemical model and provide an interpretable empirical equation for predicting the plating onset state of charge. We then explore the reversibility of lithium plating and its connection to electrolyte design for preventing irreversible Li accumulation. Finally, we design a method to quantify in situ Li plating for commercially relevant graphite|LiNi 0.5 Mn 0.3 Co 0.2 O 2 (NMC) cells and compare with results from the experimentally convenient Li|graphite configuration. The hypotheses and abundant data herein were generated primarily with equipment universal to the battery researcher, encouraging further development of innovative testing methods and data processing that enable rapid battery engineering.

25 ENERGY STORAGE↗

Feedback density and causal complexity of simulation model structure

Measures of simulation model complexity generally focus on outputs; we propose measuring the complexity of a model’s causal structure to gain insight into its fundamental character. This article introduces tools for measuring causal complexity. First, we introduce a method for developing a model’s causal structure diagram, which characterises the causal interactions present in the code. Causal structure diagrams facilitate comparison of simulation models, including those from different paradigms. Next, we develop metrics for evaluating a model’s causal complexity using its causal structure diagram. We discuss cyclomatic complexity as a measure of the intricacy of causal structure and introduce two new metrics that incorporate the concept of feedback, a fundamental component of causal structure. The first new metric introduced here is feedback density, a measure of the cycle-based interconnectedness of causal structure. The second metric combines cyclomatic complexity and feedback density into a comprehensive causal complexity measure. Finally, we demonstrate these complexity metrics on simulation models from multiple paradigms and discuss potential uses and interpretations. These tools enable direct comparison of models across paradigms and provide a mechanism for measuring and discussing complexity based on a model’s fundamental assumptions and design.

97 MATHEMATICS AND COMPUTING↗

Training custom light curve models of SN Ia subpopulations selected according to host galaxy properties

ABSTRACT Type Ia supernova (SN Ia) cosmology analyses include a luminosity step function in their distance standardization process to account for an observed yet unexplained difference in the post-standardization luminosities of SNe Ia originating from different host galaxy populations [e.g. high-mass ($M \gtrsim 10^{10} \, {\rm M}_{\odot }$) versus low-mass galaxies]. We present a novel method for including host-mass correlations in the SALT3 (Spectral Adaptive Light curve Template 3) light curve model used for standardizing SN Ia distances. We split the SALT3 training sample according to host-mass, training independent models for the low- and high-host-mass samples. Our models indicate that there are different average Si ii spectral feature strengths between the two populations, and that the average spectral energy distribution of SNe from low-mass galaxies is bluer than the high-mass counterpart. We then use our trained models to perform an SN cosmology analysis on the 3-yr spectroscopically confirmed Dark Energy Survey SN sample, treating SNe from low- and high-mass host galaxies as separate populations throughout. We find that our mass-split models reduce the Hubble residual scatter in the sample, albeit at a low statistical significance. We do find a reduction in the mass-correlated luminosity step but conclude that this arises from the model-dependent re-definition of the fiducial SN absolute magnitude rather than the models themselves. Our results stress the importance of adopting a standard definition of the SN parameters (x0, x1, c) in order to extract the most value out of the light curve modelling tools that are currently available and to correctly interpret results that are fit with different models.

Taylor, G. (ORCID:0000000157563259)↗

Using Fourier-Plane Imaging Microscopy for Determining Transition-Dipole-Moment Orientations in Organic Light-Emitting Devices

Here, we use Fourier-plane imaging microscopy (FIM) to determine the transition-dipole-moment orientation in doped organic emissive thin films. The use of FIM enables precise, sensitive, and rapid measurement of dipole orientation in the emission layer of an organic light-emitting device (OLED). An optical model of a stratified birefringent multilayer is introduced for interpreting results obtained by FIM. Using the model, we determine the average orientation of transition-dipole-moment vectors of three phosphorescent dopant emitters. The dipole alignment measured by FIM quantitatively explains the difference in OLED efficiencies using these archetype dopant molecules. FIM provides a nondestructive tool to measure and ultimately improve the outcoupling efficiency of OLEDs and other light-emitting devices.

74 ATOMIC AND MOLECULAR PHYSICS↗

PDF DECODER ANALYSIS CODE

SF-24-038"PDFdecoder", as a new application to explore parametrizations of parton distribution functions (PDFs) of the proton or other hadrons. The PDFs are fundamental quantities in particle physics which are necessary inputs to precise theoretical predictions for experiments at the Large Hadron Collider (LHC) and other facilities. As such, understanding how the PDFs are parametrized and associated uncertainties is a pressing need. The specific problem PDFdecoder confronts is the need of having a tractable and interpretably machine-learning (ML) framework to parametrize the PDFs and their uncertainties so as to understand how a given preferred parametrization is obtained. This problem has not been significantly addressed in the current literature. While other groups have used ML-based approaches to parametrize PDFs in the form of feed-forward neural networks, the question of tractability has not been explored in a PDF context. Our solution makes significant progress in this problem by using an array of encoder-decoder (essentially, autoencoder) architectures with varying constraints to the intermediate latent spaces based on interpretable physics. As a consequence, the trained models can be used as generative networks to produce interpretable predictions for the PDFs in a way that can be refined and studied further.

Hobbs, Timothy↗

Grand Unified Models

Attempts to place compact radio sources within a general interpretive framework for galactic nuclear activity, or Grand Unified Models, are reviewed. The study of superluminal motion is discussed, stressing the need for more objective methods to compare VLBI maps of compact radio sources at different epochs. Observations and arguments for and against the importance of relativistic beaming in compact radio sources are presented. The simple source model, in which the observed components move with uniform velocity along the source symmetry axis away from a stationary self-absorbed core is examined and suggestions for modifying the model are given. The parabolic spectra of blazars are discussed. The black hole model for explaining general activity in the nuclei of galaxies and radio emission and the incorporation of the apparent evolutionary properties of active galactic nuclei into grand models are also examined.

Blandford, Roger D.↗