SEARCH · Engineering Papers

Results for “explainability”

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 55 records · Page 3

All sources NASA NTRS DOE OSTI

DOE OSTI2021

Assessing Decadal Predictability in an Earth-System Model Using Explainable Neural Networks

We show that explainable neural networks can identify regions of oceanic variability that contribute predictability on decadal timescales in a fully coupled Earth-system model. The neural networks learn to use sea-surface temperature anomalies to predict future continental surface temperature anomalies. We then use a neural-network explainability method called layerwise relevance propagation to infer which oceanic patterns lead to accurate predictions made by the neural networks. In particular, regions within the North Atlantic Ocean and North Pacific Ocean lend the most predictability for surface temperature across continental North America. We apply the proposed methodology to decadal variability, although the concept is generalizable to other timescales of predictability. Furthermore, while our approach focuses on predictable patterns of internal variability within climate models, it should be generalizable to observational data as well. Our study contributes to the growing evidence that explainable neural networks are important tools for advancing geoscientific knowledge.

58 GEOSCIENCES↗

DOE OSTI2026

Violent mergers can explain the inflated state of some of the fastest stars in the Galaxy

A significant number of hypervelocity stars with velocities between 1500 − 2500 km s −1 have recently been observed. The only plausible explanation so far is that they were produced through thermonuclear supernovae in white dwarf binaries. Since these stars are thought to be surviving donors of Type Ia supernovae, a surprising finding was that these stars are inflated, with radii an order of magnitude higher than expected for Roche-lobe-filling donors. Recent attempts at explaining them have combined 3D hydrodynamical supernova explosion simulations with 1D stellar modelling to explain the impact of supernova shocks on runaway white dwarfs. However, only the hottest and most compact of those runaway stars can so far marginally be reproduced by detailed models of runaways from supernova explosions. In this and a companion paper, we introduce a new AREPO simulation of two massive CO white dwarfs that explode via a violent merger. During the merger, the primary white dwarf ignites when the secondary is on its last orbit and plunging towards the primary. In the corresponding aftermath, the core of the secondary white dwarf of 0.16 M ⊙ remains bound, moving at a velocity of ∼2800 km s −1 . We mapped this object into MESA and show that this runaway star can explain the observations of two hypervelocity stars that were dubbed D6-1 and D6-3 based on their original discovery motivated by the D6 scenario, though the violent merger scenario presented here is somewhat distinct from the D6 scenario.

Astronomy and AstroPhysics↗

DOE OSTI2026

Global Explainability of A Deep Abstaining Classifier for Cancer Pathology Reports

We present a global explainability method to characterize sources of errors in a real-world multitask deep abstaining classifier (DAC), in the context of cancer histology prediction. Our multitask classifier, currently deployed for automated annotation of cancer pathology reports from NCI-SEER registries, was trained and evaluated on 1.04 million hand-annotated samples and makes simultaneous predictions of cancer site, subsite, histology, laterality, and behavior for each report. The DAC framework enables the model to abstain on ambiguous reports and confusing classes to achieve the target accuracy on the retained (non-abstained) samples, but at the cost of decreased coverage. Requiring 97% accuracy on the histology task caused our model to retain only 22% of all samples, mostly the less ambiguous and common classes. Local explainability with the GradInp technique provided a computationally efficient way of obtaining contextual reasoning for hundreds of thousands of individual predictions. Our method, involving dimensionality reduction of approximately 13000 aggregated local explanations (ALE), offers a tractable path to true global explainability. It enabled identification of sources of errors in histology classification, globally, as hierarchical complexity among classes, label noise, insufficient information, and conflicting evidence. This suggests several strategies for iterative improvement of our DAC, including well-designed exclusion criteria, focused annotation, and reduced penalties for errors involving hierarchically related classes.

59 BASIC BIOLOGICAL SCIENCES↗

DOE OSTI2021

SAGE Intrusion Detection System: Sensitivity Analysis Guided Explainability for Machine Learning.

This report details the results of a three-fold investigation of sensitivity analysis (SA) for machine learning (ML) explainability (MLE): (1) the mathematical assessment of the fidelity of an explanation with respect to a learned ML model, (2) quantifying the trustworthiness of a prediction, and (3) the impact of MLE on the efficiency of end-users through multiple users studies. We focused on the cybersecurity domain as the data is inherently non-intuitive. As ML is being using in an increasing number of domains, including domains where being wrong can elicit high consequences, MLE has been proposed as a means of generating trust in a learned ML models by end users. However, little analysis has been performed to determine if the explanations accurately represent the target model and they themselves should be trusted beyond subjective inspection. Current state-of-the-art MLE techniques only provide a list of important features based on heuristic measures and/or make certain assumptions about the data and the model which are not representative of the real-world data and models. Further, most are designed without considering the usefulness by an end-user in a broader context. To address these issues, we present a notion of explanation fidelity based on Shapley values from cooperative game theory. We find that all of the investigated MLE explainability methods produce explanations that are incongruent with the ML model that is being explained. This is because they make critical assumptions about feature independence and linear feature interactions for computational reasons. We also find that in deployed, explanations are rarely used due to a variety of reason including that there are several other tools which are trusted more than the explanations and there is little incentive to use the explanations. In the cases when the explanations are used, we found that there is the danger that explanations persuade the end users to wrongly accept false positives and false negatives. However, ML model developers and maintainers find the explanations more useful to help ensure that the ML model does not have obvious biases. In light of these findings, we suggest a number of future directions including developing MLE methods that directly model non-linear model interactions and including design principles that take into account the usefulness of explanations to the end user. We also augment explanations with a set of trustworthiness measures that measure geometric aspects of the data to determine if the model output should be trusted.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2020

Exploring Whether Super-puffs can be Explained as Ringed Exoplanets

An intriguing, growing class of planets are the “super-puffs,” objects with exceptionally large radii for their masses and thus correspondingly low densities (≲0.3 g cm{sup −3}). Here we consider whether they could have large inferred radii because they are in fact ringed. This would naturally explain why super-puffs have thus far only shown featureless transit spectra. We find that this hypothesis can work in some cases but not all. The close proximity of the super-puffs to their parent stars necessitates rings with a rocky rather than icy composition. This limits the radius of the rings, and makes it challenging to explain the large size of Kepler 51b, 51c, 51d, and 79d unless the rings are composed of porous material. Furthermore, the short tidal locking timescales for Kepler 18d, 223d, and 223e mean that these planets may be spinning too slowly, resulting in a small oblateness and rings that are warped by their parent star. Kepler 87c and 177c have the best chance of being explained by rings. Using transit simulations, we show that testing this hypothesis requires photometry with a precision of somewhere between ∼10 ppm and ∼50 ppm, which roughly scales with the ratio of the planet and star’s radii. We conclude with a note about the recently discovered super-puff HIP 41378f.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2017

Which Bow Shock Theory, Gasdynamic or Magnetohydrodynamic, Better Explains CME Stand-off Distance Ratios from LASCO-C2 Observations ?

It is generally believed that fast coronal mass ejections (CMEs) can generate their associated shocks, which are characterized by faint structures ahead of CMEs in white-light coronagraph images. In this study, we examine whether the observational stand-off distance ratio, defined as the CME stand-off distance divided by its radius, can be explained by bow shock theories. Of 535 SOHO /LASCO CMEs (from 1996 to 2015) with speeds greater than 1000 km s{sup −1} and angular widths wider than 60°, we select 18 limb CMEs with the following conditions: (1) their Alfvénic Mach numbers are greater than one under Mann’s magnetic field and Saito’s density distributions; and (2) the shock structures ahead of the CMEs are well identified. We determine observational CME stand-off distance ratios by using brightness profiles from LASCO-C2 observations. We compare our estimates with theoretical stand-off distance ratios from gasdynamic (GD) and magnetohydrodynamic (MHD) theories. The main results are as follows. Under the GD theory, 39% (7/18) of the CMEs are explained in the acceptable ranges of adiabatic gamma ( γ ) and CME geometry. Under the MHD theory, all the events are well explained when we consider quasi-parallel MHD shocks with γ = 5/3. When we use polarized brightness (pB) measurements for coronal density distributions, we also find similar results: 8% (1/12) under GD theory and 100% (12/12) under MHD theory. Our results demonstrate that the bow shock relationships based on MHD theory are more suitable than those based on GD theory for analyzing CME-driven shock signatures.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2026

Using Explainable Artificial Intelligence to Predict Perovskite Solar Cell Electrical Metastability from Operando Photoluminescence Images in Accelerated Stress Testing

Metal halide perovskite (MHP) solar cells exhibit a metastable response to bias governed by coupled ionic–electronic processes, complicating the conventional reciprocity relation between luminescence intensity and device open-circuit voltage (V oc ). This limits the use of luminescence as a diagnostic for device screening or accelerated stress testing, motivating new approaches that can interpret photoluminescence (PL) signals under nonequilibrium conditions. From the artificial intelligence perspective, we develop an explainable deep learning framework that integrates convolutional neural networks (CNN), long short-term memory (LSTM) layers, and an attention mechanism to learn spatiotemporal features from operando photoluminescence PL image sequences. The model achieves a mean absolute error of ±0.027 V in predicting open-circuit voltage transients and reduces extreme-tail errors by up to 78% compared to physics-based reciprocity calculations. Gradient-weighted Class Activation Mapping (Grad-CAM) provides interpretability by highlighting physically meaningful regions such as electrode edges and emergent defect features. From the engineering application perspective, this framework enables accurate, contactless prediction of device V oc and identification of degradation-relevant features during accelerated aging of perovskite solar cells. This approach demonstrates how explainable AI can enhance operando diagnostics and reliability analysis in photovoltaic devices under nonequilibrium conditions.

14 SOLAR ENERGY↗

DOE OSTI2023

Explainable machine learning model for multi-step forecasting of reservoir inflow with uncertainty quantification

We propose an explainable machine learning (ML) model with uncertainty quantification (UQ) to improve multi-step reservoir inflow forecasting. Traditional ML methods have challenges in forecasting inflows multiple days ahead, and lack explainability and UQ. To address these limitations, we introduce an encoder–decoder long short-term memory (ED-LSTM) network for multi-step forecasting, employ the SHapley Additive exPlanation (SHAP) technique for understanding the influence of hydrometeorological factors on inflow prediction, and develop a novel UQ method for prediction trustworthiness. We apply these methods to forecast 7-day inflow in snow-dominant and rain-driven reservoirs. The results demonstrate the effectiveness of the ED-LSTM model, with high forecasting accuracy for short lead times. Our UQ method provides reliable uncertainty estimates, covering 90% of data with a 90% confidence level. The SHAP analysis reveals the importance of historical inflow and precipitation as influential factors. These findings and methods may support reservoir operators in optimizing water resources management decisions.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2025

Aerosol Influences on Cloud Water: Insights From ARM EPCAPE Observations With Explainable Machine Learning

This study employs an explainable machine learning (ML) framework (XGBoost-SHapley Additive exPlanations analysis) to investigate controlling factors on cloud liquid water path (LWP) using EPCAPE observations near the California coast. Aerosols are found to be the dominant factor explaining LWP variability, surpassing meteorological factors (MFs). By isolating aerosol effects from meteorological influences, the ML reveals a negative linear relationship between LWP and cloud droplet number concentration (Nd) in log space, likely driven by entrainment drying via evaporation-entrainment feedback. This aligns with the negative regime of the inverted-V relationship reported in previous studies, while no positive LWP responses are found due to a limited number of precipitating cases in EPCAPE. Furthermore, the sensitivity of LWP to Nd shows a non-linear dependence on MFs like moisture contrast between surface and free troposphere and lower-tropospheric stability. This occurs due to the interplay between the MFs' direct effects on entrainment drying and indirect effects through LWP adjustments.

EPCAPE observations↗

DOE OSTI2021

EXPLAINABLE AND TRUSTWORTHY DIAGNOSTICS ACHIEVABLE THROUGH PROCESS-BASED AUTOMATED REASONING

An approach has been developed that incorporates domain knowledge to obtain a more explainable and trustworthy equipment health monitoring diagnosis than might otherwise be obtained from a purely data-driven method. Physics-based models serve to constrain the realizable solution space and render a more trusted diagnosis. An automated reasoning algorithm performs backward chaining to infer a diagnosis that is consistent with logic statements that have been evaluated as true. This diagnosis is made explainable to an operator by providing the forward chaining path that elucidates for inspection and validity testing those truths implied by the diagnosis.

automated reasoning↗

DOE OSTI2024

TCR-H: explainable machine learning prediction of T-cell receptor epitope binding on unseen datasets

Artificial-intelligence and machine-learning (AI/ML) approaches to predicting T-cell receptor (TCR)-epitope specificity achieve high performance metrics on test datasets which include sequences that are also part of the training set but fail to generalize to test sets consisting of epitopes and TCRs that are absent from the training set, i.e., are ‘unseen’ during training of the ML model. We present TCR-H, a supervised classification Support Vector Machines model using physicochemical features trained on the largest dataset available to date using only experimentally validated non-binders as negative datapoints. TCR-H exhibits an area under the curve of the receiver-operator characteristic (AUC of ROC) of 0.87 for epitope ‘hard splitting’ (i.e., on test sets with all epitopes unseen during ML training), 0.92 for TCR hard splitting and 0.89 for ‘strict splitting’ in which neither the epitopes nor the TCRs in the test set are seen in the training data. Furthermore, we employ the SHAP (Shapley additive explanations) eXplainable AI (XAI) method for post hoc interrogation to interpret the models trained with different hard splits, shedding light on the key physiochemical features driving model predictions. TCR-H thus represents a significant step towards general applicability and explainability of epitope:TCR specificity prediction.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2025

Role of Uncertainty Quantification in the Explainability of Large Language Models for the Nuclear Industry

The meteoric rise of generative artificial intelligence (AI) large language models (LLMs) has created an opportunity to utilize them to increase efficiencies in a multitude of industries. While LLMs carry great potential to revolutionize the manner in which work is performed, numerous known deficiencies limit their utility, including the black box nature of the models, the stochastic nature of the response (i.e., presenting the same prompt multiple times results in different responses), and the potential for hallucination. Widespread adoption of LLMs in safety-critical industries such as nuclear will require some form of explainability to assure end users that the LLM’s response to a given query is valid. Model uncertainty is inherently linked to the concepts of trust and explainability, and can be used to identify situations in which the model is insufficiently certain about its answer. Although uncertainty is not enough in and of itself to determine the suitability of an answer—a model can be very certain of an inaccurate answer—it still provides valuable supporting information. Practical methodologies for gauging or quantifying the uncertainty in LLM outputs are presented herein, along with examples based on nuclear-specific prompts.

22 - GENERAL STUDIES OF NUCLEAR REACTORS↗

DOE OSTI2016

Root traits explain observed tundra vegetation nitrogen uptake patterns: Implications for trait-based land models: Tundra N Uptake Model-Data Comparison

Ongoing climate warming will likely perturb vertical distributions of nitrogen availability in tundra soils through enhancing nitrogen mineralization and releasing previously inaccessible nitrogen from frozen permafrost soil. But, arctic tundra responses to such changes are uncertain, because of a lack of vertically explicit nitrogen tracer experiments and untested hypotheses of root nitrogen uptake under the stress of microbial competition implemented in land models. We conducted a vertically explicit 15N tracer experiment for three dominant tundra species to quantify plant N uptake profiles. Then we applied a nutrient competition model (N-COM), which is being integrated into the ACME Land Model, to explain the observations. Observations using an 15 N tracer showed that plant N uptake profiles were not consistently related to root biomass density profiles, which challenges the prevailing hypothesis that root density always exerts first-order control on N uptake. By considering essential root traits (e.g., biomass distribution and nutrient uptake kinetics) with an appropriate plant-microbe nutrient competition framework, our model reasonably reproduced the observed patterns of plant N uptake. Additionally, we show that previously applied nutrient competition hypotheses in Earth System Land Models fail to explain the diverse plant N uptake profiles we observed. These results cast doubt on current climate-scale model predictions of arctic plant responses to elevated nitrogen supply under a changing climate and highlight the importance of considering essential root traits in large-scale land models. Finally, we provided suggestions and a short synthesis of data availability for future trait-based land model development.

15N tracer↗

DOE OSTI2016

Revisiting streamside trees that do not use stream water: can the two water worlds hypothesis and snowpack isotopic effects explain a missing water source?: Riparian tree water sources

We revisit a classic ecohydrological study that showed streamside riparian trees in a semiarid mountain catchment did not use perennial stream water. The original study suggested that mature individuals of Acer negundo, Acer grandidentatum, and other species were dependent on water from “deeper strata,” possibly groundwater. We used a dual stable isotope approach (δ 18 O and δ 2 H) to further examine the water sources of these trees. We tested the hypothesis that groundwater was the main tree water source, but found that neither groundwater nor stream water matched the isotope composition of xylem water during two growing seasons. Soil water (0–1 m depth) was closest to and periodically overlapped with xylem water isotope composition, but overall, xylem water was isotopically enriched compared to all measured water sources. The “two water worlds” hypothesis postulates that soil water comprises isotopically distinct mobile and less mobile pools that do not mix, potentially explaining this disparity. We further hypothesized that isotopic effects during snowpack metamorphosis impart a distinct isotope signature to the less mobile soil water that supplies summer transpiration. Depth trends in water isotopes following snowmelt were consistent with the two water worlds hypothesis, but snow metamorphic isotope effects could not explain the highly enriched xylem water. Thus, the dual isotope approach did not unambiguously determine the water source(s) of these riparian trees. Further exploration of physical, geochemical, and biological mechanisms of water isotope fractionation and partitioning is necessary to resolve these data, highlighting critical challenges in the isotopic determination of plant water sources.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2022

The channel-source hypothesis: Empirical evidence for in-channel sourcing of dissolved organic carbon to explain hysteresis in a headwater mountain stream

Catchment hydrologists have long puzzled over the question: How can catchments rapidly generate storm flows and pulses of solutes in response to storm events? Conceptual models viewing catchments as composed of discrete source areas generating flow at unique time scales and with unique chemical characteristics have been used to explain the observed changes in flow and water chemistry. Surprisingly, those conceptual models usually do not treat the stream channel as one of the potential source areas. We propose the channel-source hypothesis in which the stream itself should be considered as a potential source with the same rigour as other contributing areas. We pose this in the spirit of the scientific use of the word: a hypothesis is not a proven idea but rather a provisional supposition serving as the basis for further study. We suggest that the channel should be considered as a potential source for dissolved organic carbon (DOC). Channels store substantial amounts of organic matter, and stream ecologists have long studied stream carbon cycling. From those studies, we know that leaching and decomposition can generate DOC from particulate organic carbon (POC). Further, POC is stored in channel ‘dead-zones’—regions of low flow velocity - that can be activated as flow velocity increases, thus releasing accumulated DOC during storms. All catchments are different; there is no reason to assume that channel sources are always important, in every catchment, in every storm. Thus, the channel-source hypothesis does not replace existing conceptual models. Instead, it adds another potential mechanism that may explain DOC dynamics observed in streams. The channel-source hypothesis has substantial implications for catchment studies examining sources of DOC in stream water or using DOC as a tracer to determine the locations of, and proportional contributions of, different source areas for streamflow generation.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2021

Combination of deep neural network with attention mechanism enhances the explainability of protein contact prediction

Abstract Deep learning has emerged as a revolutionary technology for protein residue‐residue contact prediction since the 2012 CASP10 competition. Considerable advancements in the predictive power of the deep learning‐based contact predictions have been achieved since then. However, little effort has been put into interpreting the black‐box deep learning methods. Algorithms that can interpret the relationship between predicted contact maps and the internal mechanism of the deep learning architectures are needed to explore the essential components of contact inference and improve their explainability. In this study, we present an attention‐based convolutional neural network for protein contact prediction, which consists of two attention mechanism‐based modules: sequence attention and regional attention. Our benchmark results on the CASP13 free‐modeling targets demonstrate that the two attention modules added on top of existing typical deep learning models exhibit a complementary effect that contributes to prediction improvements. More importantly, the inclusion of the attention mechanism provides interpretable patterns that contain useful insights into the key fold‐determining residues in proteins. We expect the attention‐based model can provide a reliable and practically interpretable technique that helps break the current bottlenecks in explaining deep neural networks for contact prediction. The source code of our method is available at https://github.com/jianlin-cheng/InterpretContactMap .

59 BASIC BIOLOGICAL SCIENCES↗

DOE OSTI2015

Reducing conditions, reactive metals, and their interactions can explain spatial patterns of surface soil carbon in a humid tropical forest

Humid tropical forests support large stocks of surface soil carbon (C) that exhibit high spatial variability over scales of meters to landscapes (km). Reactive minerals and organo-metal complexes are known to contribute to C accumulation in these ecosystems, although potential interactions with environmental factors such as oxygen (O 2 ) availability have received much less attention. Reducing conditions can potentially contribute to C accumulation, yet anaerobic metabolic processes such as iron (Fe) reduction can also drive substantial C losses. We tested whether these factors could explain variation in soil C (0–10 and 10–20 cm depths) over multiple spatial scales in the Luquillo Experimental Forest, Puerto Rico, using reduced iron (Fe(II)) concentrations as an index of reducing conditions across sites differing in vegetation, topographic position, and/or climate. Fine root biomass and Fe(II) were the best overall correlates of site (n = 6) mean C concentrations and stocks from 0 to 20 cm depth (r = 0.99 and 0.98, respectively). Litterfall decreased as reducing conditions, total and dead fine root biomass, and soil C increased among sites, suggesting that decomposition rates rather than C inputs regulated soil C content at the landscape scale. Strong relationships between Fe(II) and dead fine root biomass suggest that reducing conditions suppressed particulate organic matter decomposition. The optimal mixed-effects regression model for individual soil samples (n = 149) showed that aluminum (Al) and Fe in citrate/ascorbate and oxalate extractions, Fe(II), fine root biomass, and interactions between Fe(II) and Al explained most of the variation in C concentrations (pseudo R 2 = 0.82). The optimal model of C stocks was similar but did not include fine root biomass (pseudo R 2 = 0.62). In these models, soil C concentrations and stocks increased with citrate/ascorbate-extractable Al and oxalate-extractable Fe. However, soil C decreased with citrate/ascorbate-extractable Fe, an index of Fe susceptible to anaerobic microbial reduction. At the site scale (n = 6), ratios of citrate/ascorbate to oxalate-extractable Fe consistently decreased across a landscape O 2 gradient as C increased. Finally, we suggest that the impact of reducing conditions on organic matter decomposition and the presence of organo-metal complexes and C sorption by short-range order Fe and Al contribute to C accumulation, whereas the availability of an Fe pool to sustain anaerobic respiration in soil microsites partially attenuates soil C accumulation in these ecosystems.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2025

Supercooled phase transitions in conformal dark sectors explain NANOGrav data

According to recent lore, it is difficult to explain the evidence for a stochastic gravitational wave background obtained by pulsar timing arrays with supercooled first-order phase transitions (FOPTs). We demonstrate that supercooled FOPTs in dark U(1) models with a conformal dark sector easily explain the nHz signal at NANOGrav.

Conformal models↗