Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “explainable”

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 109 records · Page 6
All sourcesNASA NTRSDOE OSTI
DOE OSTI2023

Explainable deep learning for insights in El Niño and river flows

The El Niño Southern Oscillation (ENSO) is a semi-periodic fluctuation in sea surface temperature (SST) over the tropical central and eastern Pacific Ocean that influences interannual variability in regional hydrology across the world through long-range dependence or teleconnections. Recent research has demonstrated the value of Deep Learning (DL) methods for improving ENSO prediction as well as Complex Networks (CN) for understanding teleconnections. However, gaps in predictive understanding of ENSO-driven river flows include the black box nature of DL, the use of simple ENSO indices to describe a complex phenomenon and translating DL-based ENSO predictions to river flow predictions. Here we show that eXplainable DL (XDL) methods, based on saliency maps, can extract interpretable predictive information contained in global SST and discover SST information regions and dependence structures relevant for river flows which, in tandem with climate network constructions, enable improved predictive understanding. Our results reveal additional information content in global SST beyond ENSO indices, develop understanding of how SSTs influence river flows, and generate improved river flow prediction, including uncertainty estimation. Observations, reanalysis data, and earth system model simulations are used to demonstrate the value of the XDL-CN based methods for future interannual and decadal scale climate projections.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2025

Photolytic activation of Ni (II) X 2 L explains how Ni-mediated cross coupling begins

Nickel photocatalysis has recently become vital to organic synthesis, but how the Ni (II) X 2 L pre-catalyst (X = Cl, Br; L = bidentate ligand) becomes activated to Ni (I) XL has remained puzzling and is typically addressed on a case-by-case basis. Here, we reveal a general mechanism where light induces photolysis of the Ni (II) -X bond, either via direct excitation or triplet energy transfer. Photolysis produces Ni (I) XL and a halogen radical, X*. Subsequent hydrogen atom abstraction, often from the solvent, produces a C(sp 3 ) radical, R*, that recombines with Ni (I) to form organonickel(II) complexes, Ni (II) XRL. Rather than acting as a loss pathway, Ni (II) XRL behaves as a light-activated reservoir of Ni (I) via photolysis of the Ni (II) -C bond. These results explain the role of the solvent in protecting the catalyst from off-cycle dimerization, demonstrate that two photons are often required to drive the reaction, and show how tuning the ligand can control the concentration of active Ni (I) species.

08 HYDROGEN↗
DOE OSTI2022

Extracting structural motifs from pair distribution function data of nanostructures using explainable machine learning

Characterization of material structure with X-ray or neutron scattering using e.g. Pair Distribution Function (PDF) analysis most often rely on refining a structure model against an experimental dataset. However, identifying a suitable model is often a bottleneck. Recently, automated approaches have made it possible to test thousands of models for each dataset, but these methods are computationally expensive and analysing the output, i.e. extracting structural information from the resulting fits in a meaningful way, is challenging. Our Machine Learning based Motif Extractor (ML-MotEx) trains an ML algorithm on thousands of fits, and uses SHAP (SHapley Additive exPlanation) values to identify which model features are important for the fit quality. We use the method for 4 different chemical systems, including disordered nanomaterials and clusters. ML-MotEx opens for a type of modelling where each feature in a model is assigned an importance value for the fit quality based on explainable ML.

36 MATERIALS SCIENCE↗
DOE OSTI2023

Deep material network via a quilting strategy: visualization for explainability and recursive training for improved accuracy

Recent developments integrating micromechanics and neural networks offer promising paths for rapid predictions of the response of heterogeneous materials with similar accuracy as direct numerical simulations. The deep material network is one such approaches, featuring a multi-layer network and micromechanics building blocks trained on anisotropic linear elastic properties. Once trained, the network acts as a reduced-order model, which can extrapolate the material’s behavior to more general constitutive laws, including nonlinear behaviors, without the need to be retrained. However, current training methods initialize network parameters randomly, incurring inevitable training and calibration errors. Here, we introduce a way to visualize the network parameters as an analogous unit cell and use this visualization to “quilt” patches of shallower networks to initialize deeper networks for a recursive training strategy. The result is an improvement in the accuracy and calibration performance of the network and an intuitive visual representation of the network for better explainability.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Global burned area increasingly explained by climate change

Fire behaviour is changing in many regions worldwide. However, nonlinear interactions between fire weather, fuel, land use, management and ignitions have impeded formal attribution of global burned area changes. Here, in this work, we demonstrate that climate change increasingly explains regional burned area patterns, using an ensemble of global fire models. The simulations show that climate change increased global burned area by 15.8% (95% confidence interval (CI) [13.1–18.7]) for 2003–2019 and increased the probability of experiencing months with above-average global burned area by 22% (95% CI [18–26]). In contrast, other human forcings contributed to lowering burned area by 19.1% (95% CI [21.9–15.8]) over the same period. Moreover, the contribution of climate change to burned area increased by 0.22% (95% CI [0.22–0.24]) per year globally, with the largest increase in central Australia. Our results highlight the importance of immediate, drastic and sustained GHG emission reductions along with landscape and fire management strategies to stabilize fire impacts on lives, livelihoods and ecosystems.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2021

Root traits explain plant species distributions along climatic gradients yet challenge the nature of ecological trade-offs

Ecological theory is built on trade-offs, where trait differences among species evolved as adaptations to different environments. Trade-offs are often assumed to be bidirectional, where opposite ends of a gradient in trait values confer advantages in different environments. However, unidirectional benefits could be widespread if extreme trait values confer advantages at one end of an environmental gradient, whereas a wide range of trait values are equally beneficial at the other end. Here, we show that root traits explain species occurrences along broad gradients of temperature and water availability, but model predictions only resembled trade-offs in two out of 24 models. Forest species with low specific root length and high root tissue density (RTD) were more likely to occur in warm climates but species with high specific root length and low RTD were more likely to occur in cold climates. Unidirectional benefits were more prevalent than trade-offs: for example, species with large-diameter roots and high RTD were more commonly associated with dry climates, but species with the opposite trait values were not associated with wet climates. Directional selection for traits consistently occurred in cold or dry climates, whereas a diversity of root trait values were equally viable in warm or wet climates. Explicit integration of unidirectional benefits into ecological theory is needed to advance our understanding of the consequences of trait variation on species responses to environmental change.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2022

A simple model of COVID-19 explains disease severity and the effect of treatments

Abstract Considerable effort has been made to better understand why some people suffer from severe COVID-19 while others remain asymptomatic. This has led to important clinical findings; people with severe COVID-19 generally experience persistently high levels of inflammation, slower viral load decay, display a dysregulated type-I interferon response, have less active natural killer cells and increased levels of neutrophil extracellular traps. How these findings are connected to the pathogenesis of COVID-19 remains unclear. We propose a mathematical model that sheds light on this issue by focusing on cells that trigger inflammation through molecular patterns: infected cells carrying pathogen-associated molecular patterns (PAMPs) and damaged cells producing damage-associated molecular patterns (DAMPs). The former signals the presence of pathogens while the latter signals danger such as hypoxia or lack of nutrients. Analyses show that SARS-CoV-2 infections can lead to a self-perpetuating feedback loop between DAMP expressing cells and inflammation, identifying the inability to quickly clear PAMPs and DAMPs as the main contributor to hyperinflammation. The model explains clinical findings and reveal conditions that can increase the likelihood of desired clinical outcome from treatment administration. In particular, the analysis suggest that antivirals need to be administered early during infection to have an impact on disease severity. The simplicity of the model and its high level of consistency with clinical findings motivate its use for the formulation of new treatment strategies.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2025

Changes in sea ice concentration explain half of the winter warming of the Arctic surface

Arctic winter warming is stronger than in summer, but its driving mechanisms remain debated, particularly the roles of local processes, like sea-ice loss, versus remote factors, like atmospheric heat transport. Here we introduce a novel decomposition framework that characterizes Arctic warming as a function of historical atmospheric circulation, sea ice concentration, and carbon dioxide changes using observational and reanalysis data. We show that sea ice changes explain about 55% of the winter Arctic near-surface temperature trend during 1959–2015, after removing the effects directly connected to atmospheric circulation. Dynamically induced warming accounts for about 20% at surface and up to 80% in mid-troposphere. The remaining ~25% is attributed to the increase in carbon dioxide, though it also indirectly affects sea-ice loss and circulation-related warming. These findings highlight the dominant role of sea ice loss and change in atmospheric dynamics in affecting the historical Arctic winter warming.

Arctic Sea Ice change↗
DOE OSTI2022

Explaining the structure sensitivity of Pt and Rh for aqueous-phase hydrogenation of phenol

Phenol is an important model compound to understand the thermocatalytic (TCH) and electrocatalytic hydrogenation (ECH) of biomass to biofuels. Although Pt and Rh are among the most studied catalysts for aqueous-phase phenol hydrogenation, the reason why certain facets are active for ECH and TCH is not fully understood. Herein, we identify the active facet of Pt and Rh catalysts for aqueous-phase hydrogenation of phenol and explain the origin of the size-dependent activity trends of Pt and Rh nanoparticles. Phenol adsorption energies extracted on the active sites of Pt and Rh nanoparticles on carbon by fitting kinetic data show that the active sites adsorb phenol weakly. We predict that the turnover frequencies (TOFs) for the hydrogenation of phenol to cyclohexanone on Pt(111) and Rh(111) terraces are higher than those on (221) stepped facets based on density functional theory modeling and mean-field microkinetic simulations. The higher activities of the (111) terraces are due to lower activation energies and weaker phenol adsorption, preventing high coverages of phenol from inhibiting hydrogen adsorption. We measure that the TOF for ECH of phenol increases as the Rh nanoparticle diameter increases from 2 to 10 nm at 298 K and -0.1 V vs the reversible hydrogen electrode, qualitatively matching prior reports for Pt nanoparticles. The increase in experimental TOFs as Pt and Rh nanoparticle diameters increase is due to a larger fraction of terraces on larger particles. These findings clarify the structure sensitivity and active site of Pt and Rh for the hydrogenation of phenol and will inform the catalyst design for the hydrogenation of bio-oils.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2026

Attention-based explainability for structure–property relationships

Machine learning methods are emerging as a universal paradigm for constructing correlative structure–property relationships in materials science based on multimodal characterization. However, this necessitates the development of methods for the physical interpretability of the resulting correlative models. Here, we demonstrate the potential of attention-based neural networks for revealing structure–property relationships and the underlying physical mechanisms, using the ferroelectric properties of PbTiO3 thin films as a case study. Through the analysis of attention scores, we disentangle the influence of distinct domain patterns on the polarization switching process. The attention-based Transformer model is explored both as a direct interpretability tool and as a surrogate for explaining representations learned via unsupervised machine learning, enabling the identification of physically grounded correlations. We compare attention-derived interpretability scores with classical SHapley Additive exPlanations analysis and show that, in contrast to applications in natural language processing, attention mechanisms in materials science exhibit high efficiency in highlighting meaningful structural features.

Slautin, Boris [Independent Researcher]↗
DOE OSTI2021

Magnetically driven short-range order can explain anomalous measurements in CrCoNi

Significance High-entropy alloys, an exciting new class of structural materials involving many concentrated elements, are generally assumed to be chemically disordered. This work reveals how magnetic interactions can drive atomic-scale ordering in a mechanically interesting system that has been widely studied as a model for many promising alloys. The formation of short-range order in this manner is used to explain experimental anomalies, suggesting that many related materials could contain similarly impactful ordering under standard conditions. This possibility offers solutions to outstanding questions concerning CrCoNi and emphasizes the importance of magnetism and short-range order in related high-entropy alloys.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Competition response of cloud supersaturation explains diminished Twomey effect for smoky aerosol in the tropical Atlantic

The Twomey effect brightens clouds by increasing aerosol concentrations, which activates more droplets and decreases cloud supersaturation in response to more competition for water vapor. To quantify this competition response, we used marine low cloud observations in clean and smoky conditions at Ascension Island in the tropical South Atlantic during the Layered Aerosol Smoke Interactions with Cloud (LASIC) campaign. These observations show similar increases in droplet number for increased accumulation-mode particles from surface-based and satellite cloud retrievals, demonstrating the importance of below-cloud aerosol measurements for retrieving aerosol–cloud interactions (ACI) in clean and smoky aerosol conditions. Four methods for estimating cloud supersaturation from aerosol–cloud measurements were compared, with cloud scene-based and parcel-based methods showing sufficient variability for a strong dependence on both aerosol accumulation number concentration and cloud-base updraft velocities. Decomposing aerosol-related changes in cloud albedo and optical depth shows the calculated competition response accounts for dampening the activation response by 12 to 35%, explaining the diminished Twomey effect at high aerosol concentrations observed for smoky conditions at LASIC and previously around the world. This result was consistent for independent supersaturation retrievals by cloud scene-based droplet number and cloud condensation nuclei and parcel-based multimode size-resolving Lagrangian methods. Translating aerosol effects to local radiative forcing with clean conditions as a proxy for preindustrial and smoky conditions for present-day showed that the competition response reduces cooling from the Twomey radiative forcing by 12 to 35%, providing an essential process-specific constraint for improving the representation of aerosol competition in climate model simulation of indirect aerosol forcing.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2025

Observed declines in leaf nitrogen explained by photosynthetic acclimation to CO 2

Widespread evidence of decreasing leaf nutrients has raised concerns about ecosystem productivity under global change. Interpreting trends in leaf nutrients has important implications for the fate of ecosystem services, particularly the role of forests in mitigating climate change and sustaining quality food sources. Here, we challenge the common interpretation that decreasing leaf nitrogen concentration (LNC) is evidence of increasing nutrient limitations on ecosystem primary productivity. Instead, we show that declines in LNC (4% decrease per 50 ppm CO 2 increase), observed across 409 European forest plots over 22 y, can be explained by reduced photosynthetic nitrogen demand. This regional trend is consistent with leaf acclimation to increasing atmospheric CO 2 according to optimality theory. This finding suggests that enhanced photosynthetic nitrogen use efficiency due to CO 2 fertilization may lead to less nitrogen uptake and/or reallocation of nitrogen for plant growth and other functions. Our results have large implications for understanding and simulating interactions between ecosystem nitrogen and carbon cycles and suggest nitrogen requirements for terrestrial carbon uptake under elevated CO 2 may be lower than previously thought.

CO2 fertilization↗
DOE OSTI2024

The magnetic gradient scale length explains why certain plasmas require close external magnetic coils

Abstract The separation between the last closed flux surface of a plasma and the external coils that magnetically confine it is a limiting factor in the construction of fusion-capable plasma devices. This plasma-coil separation must be large enough so that components such as a breeding blanket and neutron shielding can fit between the plasma and the coils. Plasma-coil separation affects reactor size, engineering complexity, and particle loss due to field ripple. For some plasmas it can be difficult to produce the desired flux surface shaping with distant coils, and for other plasmas it is infeasible altogether. Here, we seek to understand the underlying physics that limits plasma-coil separation and explain why some configurations require close external coils. In this paper, we explore the hypothesis that the limiting plasma-coil separation is set by the shortest scale length of the magnetic field as expressed by the ∇ B tensor. We tested this hypothesis on a database of > 40 stellarator and tokamak configurations. Within this database, the coil-to-plasma distance compared to the minor radius varies by over an order of magnitude. The magnetic scale length is well correlated to the coil-to-plasma distance of actual coil designs generated using the REGCOIL method (Landreman 2017 Nucl. Fusion 57 046003). Additionally, this correlation reveals a general trend that larger plasma-coil separation is possible with a small number of field periods.

Physics↗
DOE OSTI2023

Quantum biological insights into CRISPR-Cas9 sgRNA efficiency from explainable-AI driven feature engineering

Abstract CRISPR-Cas9 tools have transformed genetic manipulation capabilities in the laboratory. Empirical rules-of-thumb have been developed for only a narrow range of model organisms, and mechanistic underpinnings for sgRNA efficiency remain poorly understood. This work establishes a novel feature set and new public resource, produced with quantum chemical tensors, for interpreting and predicting sgRNA efficiency. Feature engineering for sgRNA efficiency is performed using an explainable-artificial intelligence model: iterative Random Forest (iRF). By encoding quantitative attributes of position-specific sequences for Escherichia coli sgRNAs, we identify important traits for sgRNA design in bacterial species. Additionally, we show that expanding positional encoding to quantum descriptors of base-pair, dimer, trimer, and tetramer sequences captures intricate interactions in local and neighboring nucleotides of the target DNA. These features highlight variation in CRISPR-Cas9 sgRNA dynamics between E. coli and H. sapiens genomes. These novel encodings of sgRNAs enhance our understanding of the elaborate quantum biological processes involved in CRISPR-Cas9 machinery.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2025

Explaining the extra crystal-field mode in 𝐴 ⁢Ce ⁢𝑋 2 (𝐴 = K, Rb, Na; 𝑋 = O, S, Se, Te)

A growing list of Ce-based magnets have shown an extra and heretofore unexplained crystal electric field (CEF) mode at high energies. We describe a process whereby an optical phonon can produce a split CEF mode well above the phonon energy. We use density functional theory and point-charge model calculations to estimate the phonon distortions and coupling to model this effect in KCeO 2 , showing that it accounts for the extra CEF mode observed. Furthermore, this mechanism is generic and may explain the extra modes observed on a variety of Ce 3+ compounds.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Explaining Snowball-in-Hell Phenomena in Heavy-Ion Collisions Using a Novel Thermodynamic Variable

A loosely bound hadronic molecule produced by a relativistic heavy-ion collision has been described as a “snowball in hell” since it emerges from a hadron resonance gas whose temperature is orders of magnitude larger than the binding energy of the molecule. This remarkable phenomenon can be explained in terms of a novel thermodynamic variable called the “contact” that is conjugate to the binding momentum of the molecule. The production rate of the molecule can be expressed in terms of the contact density at the kinetic freeze-out of the hadron resonance gas. It approaches a nonzero limit as the binding energy goes to 0.

Relativistic heavy-ion collisions↗
DOE OSTI2024

Explaining Neural Spike Activity for Simulated Bio-plausible Network through Deep Sequence Learning

With significant improvements in large-scale simulations of brain models, there is a growing need to develop tools for rapid analysis and interpreting the simulation results. In this work, we explore the potential of sequential deep learning models to understand and explain the network dynamics among the neurons extracted from a large-scale neural simulation in STACS (Simulation Tool for Asynchronous Cortical Stream). Our method employs a representative neuroscience model that abstracts the cortical dynamics with a reservoir of randomly connected spiking neurons with a low stable spike firing rate throughout the simulation duration. We subsequently analyze the spike dynamics of the simulated spiking neural network through an autoencoder model and an attention-based mechanism.

Kulkarni, Shruti↗