SEARCH · Engineering Papers

Results for “interpretable models”

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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 199 records · Page 11

All sources NASA NTRS DOE OSTI

DOE OSTI2023

Differentiable modelling to unify machine learning and physical models for geosciences

Process-based modelling offers interpretability and physical consistency in many domains of geosciences but struggles to leverage large datasets efficiently. Machine-learning methods, especially deep networks, have strong predictive skills yet are unable to answer specific scientific questions. Here, in this Perspective, we explore differentiable modelling as a pathway to dissolve the perceived barrier between process-based modelling and machine learning in the geosciences and demonstrate its potential with examples from hydrological modelling. ‘Differentiable’ refers to accurately and efficiently calculating gradients with respect to model variables or parameters, enabling the discovery of high-dimensional unknown relationships. Differentiable modelling involves connecting (flexible amounts of) prior physical knowledge to neural networks, pushing the boundary of physics-informed machine learning. It offers better interpretability, generalizability, and extrapolation capabilities than purely data-driven machine learning, achieving a similar level of accuracy while requiring less training data. Additionally, the performance and efficiency of differentiable models scale well with increasing data volumes. Under data-scarce scenarios, differentiable models have outperformed machine-learning models in producing short-term dynamics and decadal-scale trends owing to the imposed physical constraints. Differentiable modelling approaches are primed to enable geoscientists to ask questions, test hypotheses, and discover unrecognized physical relationships. Future work should address computational challenges, reduce uncertainty, and verify the physical significance of outputs.

58 GEOSCIENCES↗

DOE OSTI2023

A search for new resonances in multiple final states with a high transverse momentum Z boson in $\sqrt{s} $ = 13 TeV pp collisions with the ATLAS detector

A generic search for resonances is performed with events containing a Z boson with transverse momentum greater than 100 GeV, decaying into e + e – or μ + μ – . The analysed data collected with the ATLAS detector in proton-proton collisions at a centre-of-mass energy of 13 TeV at the Large Hadron Collider correspond to an integrated luminosity of 139 fb –1 . Two invariant mass distributions are examined for a localised excess relative to the expected Standard Model background in six independent event categories (and their inclusive sum) to increase the sensitivity. No significant excess is observed. Exclusion limits at 95% confidence level are derived for two cases: a model-independent interpretation of Gaussian-shaped resonances with the mass width between 3% and 10% of the resonance mass, and a specific heavy vector triplet model with the decay mode W' → ZW → ℓℓqq.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

DOE OSTI2023

maslov-group/FUN-PROSE

mRNA levels of all genes in a genome is a critical piece of information defining the overall state of the cell in a given environmental condition. Being able to reconstruct such condition-specific expression in fungal genomes is particularly important to metabolically engineer these organisms to produce desired chemicals in industrially scalable conditions. Most previous deep learning approaches focused on predicting the average expression levels of a gene based on its promoter sequence, ignoring its variation across different conditions. Here we present FUN-PROSE—a deep learning model trained to predict differential expression of individual genes across various conditions using their promoter sequences and expression levels of all transcription factors. We train and test our model on three fungal species and get the correlation between predicted and observed condition-specific gene expression as high as 0.85. We then interpret our model to extract promoter sequence motifs responsible for variable expression of individual genes. We also carried out input feature importance analysis to connect individual transcription factors to their gene targets. A sizeable fraction of both sequence motifs and TF-gene interactions learned by our model agree with previously known biological information, while the rest corresponds to either novel biological facts or indirect correlations.

Liu, Simon↗

DOE OSTI2023

FUN-PROSE: A deep learning approach to predict condition-specific gene expression in fungi

59 BASIC BIOLOGICAL SCIENCES↗

DOE OSTI2026

Using Machine Learning to Understand Electric and Hybrid Vehicles Ownership in Burdened and Nonburdened Communities

Transitioning to electric and hybrid vehicles (EHVs) for all communities is a pivotal step toward sustainable transportation and environmental conservation. This paper aims to understand the adoption of EHVs, focusing on burdened communities (BCs) in the United States. The EHV ownership-based analysis combines two datasets—behavioral data from the Puget Sound Regional Travel Survey integrated with BCs (Justice40) data covering transportation insecurity, environmental burden, social vulnerability, health vulnerability, and climate and disaster risk burden. After creating this unique database, descriptive analysis and modeling are used to analyze the data and predict EHV ownership in the future. Specifically, we use a new method that combines particle swarm optimization (PSO) with a stacking model named PSO-Stacking, which incorporates heterogeneous base learners of machine learning and deep learning. PSO applies a customized objective function to select the optimal hyperparameters for heterogeneous learners within the stacking model, effectively addressing challenges such as multicollinearity, data imbalance, nonlinearity, and overfitting. The proposed solution covers more accurate results than standard benchmark models for EHV ownership in BCs and non-BCs. In addition, the results of the PSO-Stacking method are explained using the local interpretable model-agnostic explanations technique. Results show a negative correlation between the BCs indicators, that is, higher transportation insecurity associated with lower EHV ownership. Furthermore, BCs have higher future climate risk scores, diesel particulate matter levels, and PM2.5 in the air than non-BCs because of higher conventional vehicle ownership. These communities are at higher risk and can benefit from electrification, EV infrastructure, and EV policies to address environmental challenges.

Aslam, Zeeshan [ORNL]↗

DOE OSTI2026

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↗

DOE OSTI2022

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↗

DOE OSTI2022

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↗

DOE OSTI2023

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↗

DOE OSTI2023

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↗

DOE OSTI2023

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↗

DOE OSTI2023

SULI Oral Presentation

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

DOE OSTI2025

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↗

DOE OSTI2024

Search for supersymmetry using vector boson fusion signatures and missing transverse momentum in pp collisions at $\sqrt{s}$ = 13 TeV with the ATLAS detector

This paper presents a search for supersymmetric particles in models with highly compressed mass spectra, in events consistent with being produced through vector boson fusion. The search uses 140 fb −1 of proton-proton collision data at $\sqrt{s}$ = 13 TeV collected by the ATLAS experiment at the Large Hadron Collider. Events containing at least two jets with a large gap in pseudorapidity, large missing transverse momentum, and no reconstructed leptons are selected. A boosted decision tree is used to separate events consistent with the production of supersymmetric particles from those due to Standard Model backgrounds. The data are found to be consistent with Standard Model predictions. The results are interpreted using simplified models of R-parity-conserving supersymmetry in which the lightest supersymmetric partner is a bino-like neutralino with a mass similar to that of the lightest chargino and second-to-lightest neutralino, both of which are wino-like. Lower limits at 95% confidence level on the masses of next-to-lightest supersymmetric partners in this simplified model are established between 117 and 120 GeV when the lightest supersymmetric partners are within 1 GeV in mass.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

DOE OSTI2023

Time-dependent SOLPS-ITER simulations of the tokamak plasma boundary for model predictive control using SINDy *

Abstract Time-dependent SOLPS-ITER simulations have been used to identify reduced models with the sparse identification of nonlinear dynamics (SINDy) method and develop model-predictive control of the boundary plasma state using main ion gas puff actuation. A series of gas actuation sequences are input into SOLPS-ITER to produce a dynamic response in upstream and divertor plasma quantities. The SINDy method is applied to identify reduced linear and nonlinear models for the electron density at the outboard midplane n e , s e p O M P and the electron temperature at the outer divertor T e , s e p d i v . Note that T e , s e p d i v is not necessarily the peak value of T e along the divertor. The identified reduced models are interpretable by construction (i.e. not black box), and have the form of coupled ordinary differential equations. Despite significant noise in T e , s e p d i v , the reduced models can be used to predict the response over a range of actuation levels to a maximum deviation of 0.5% in n e , s e p O M P and 5%–10% in T e , s e p d i v for the cases considered. Model retraining using time history data triggered by a preset error threshold is also demonstrated. A model predictive control strategy for nonlinear models is developed and used to perform feedback control of a SOLPS-ITER simulation to produce a setpoint trajectory in n e , s e p O M P using the integrated plasma simulator framework. The developed techniques are general and can be applied to time-dependent data from other boundary simulations or experimental data. Ongoing work is extending the approach to model identification and control for divertor detachment, which will present transient nonlinear behavior from impurity seeding, including realistic latency and synthetic diagnostic signals derived from the full SOLPS-ITER output.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

DOE OSTI2022

Benchmark eA generator for leptoproduction in high-energy lepton-nucleus collisions

The upcoming electron-ion collider (EIC) will address several outstanding puzzles in modern nuclear physics. Topics such as the partonic structure of nucleons and nuclei, the origin of their mass and spin, among others, can be understood via the study of high-energy electron-proton (ep) and electron-nucleus (eA) collisions. Achieving the scientific goals of the EIC will require a novel electron-hadron collider and detectors capable to perform high-precision measurements but also dedicated tools to model and interpret the data. To aid in the latter, we present a general-purpose e A Monte Carlo generator—BeAGLE. In this paper, we provide a general description of the models integrated into BeAGLE, applications of BeAGLE in eA physics, implications for detector requirements at the EIC, and the tuning of the parameters in BeAGLE based on available experimental data. Specifically, we focus on a selection of model and data comparisons in particle production in both ep and eA collisions, where baseline particle distributions provide essential information to characterize the event. In addition, we investigate the collision geometry determination in eA collisions, which could be used as an experimental tool for varying the nuclear density.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

DOE OSTI2020

Search for top squarks in events with a Higgs or Z boson using 139 fb -1 of pp collision data at $\sqrt{s}=13$ TeV with the ATLAS detector

This paper presents a search for direct top squark pair production in events with missing transverse momentum plus either a pair of jets consistent with Standard Model Higgs boson decay into b-quarks or a same-flavour opposite-sign dilepton pair with an invariant mass consistent with a Z boson. The analysis is performed using the proton–proton collision data at √s=13 TeV collected with the ATLAS detector during the LHC Run-2, corresponding to an integrated luminosity of 139 fb -1 . No excess is observed in the data above the Standard Model predictions. The results are interpreted in simplified models featuring direct production of pairs of either the lighter top squark ($\tilde{t}_1$) or the heavier top squark ($\tilde{t}_2$), excluding at 95% confidence level $\tilde{t}_1$ and $\tilde{t}_2$ masses up to about 1220 and 875 GeV, respectively.

42 ENGINEERING↗

DOE OSTI2026

The Role of Nuclear Data Sensitivities in Prompt α-Eigenvalue Predictions of Delayed Critical Benchmarks

Alpha (α) eigenvalues, which describe the logarithmic time derivative of the neutron population in a multiplying system, are integral to time-dependent behavior and diagnostic applications. However, uncertainties in the evaluated nuclear data can significantly impact the accuracy of transport simulations for such quantities. This work explores the use of machine learning models to predict two key outputs, α-eigenvalues and keff bias, using input features derived from α-eigenvalue sensitivities to nuclear data. The criticality safety benchmark models used in this study come from the International Handbook of Evaluated Criticality Safety Benchmark Experiments. Three models, random forest, XGBoost, and NGBoost, are trained on both energy-resolved and energy-summed α sensitivities. For the α-eigenvalue bias prediction, NGBoost achieved the highest R 2 (0.9476) using energy-resolved features, while XGBoost performed best using summed sensitivities. In contrast, when predicting the keff bias, all the models showed moderate predictive capability (best R 2 ≈ 0.72), as the mapping from the static α-sensitivities to the static keff bias was less direct. SHAP (SHapley Additive exPlanations) analysis was used to interpret the model predictions. Across both prediction tasks, the features associated with neutron capture [H-1 (n, γ)], uranium scattering reactions (such as 235 U elastic/inelastic), and actinide capture/fission reactions (such as 239 Pu and 234 U) were consistently identified as the most impactful. This highlights the key role of specific nuclear reactions and energy ranges in shaping both time-dependent and steady-state criticality behavior. These results demonstrated that α-sensitivities, despite being computed for time-dependent metrics, can provide valuable insights for predicting both α-eigenvalues and the keff bias. Moreover, machine learning models offer a promising pathway for uncovering important nuclear data dependencies and guiding future data evaluation efforts.

Nuclear data↗