Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “interpretable machine learning”

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 415 records · Page 23
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Multicode benchmark on simulated Ti K-edge x-ray absorption spectra of Ti-O compounds

X-ray absorption spectroscopy (XAS) is an element-specific materials characterization technique that is sensitive to structural and electronic properties. First-principles simulated XAS has been widely used as a powerful tool to interpret experimental spectra and draw physical insights. Recently, there has also been growing interest in building computational XAS databases to enable data analytics and machine learning applications. However, there are non-trivial differences among commonly used XAS simulation codes, both in underlying theoretical formalism and in technical implementation. Reliable and reproducible computational XAS databases require systematic benchmark studies. In this work, we benchmarked Ti K-edge XAS simulations of ten representative Ti-O binary compounds, which we refer to as the Ti-O-10 dataset, using three state-of-the-art codes: XSPECTRA, OCEAN and exciting. We systematically studied the convergence behavior with respect to the input parameters and developed a workflow to automate and standardize the calculations to ensure converged spectra. Our benchmark comparison considers a 35 eV spectral range starting from the K-edge onset, representative of widely used near-edge spectra. Quantitative comparison over this range is based on Spearman’s rank correlation score (r sp ). Our results show: (1) the two Bethe-Salpeter equation (BSE) codes (OCEAN and exciting) have excellent agreement with an average $r$ $sp$ of 0.998; (2) good agreement is obtained between the core-hole potential code (XSPECTRA) and BSE codes (OCEAN and exciting) with an average $r$ $sp$ of 0.990, and this smaller $r$ $sp$ reflects the noticeable differences in the main edge spectral shape that can be primarily attributed to the difference in the strength of the screened core-hole potential; (3) simulations from both methods overall reproduce well the main experimental spectral features of rutile and anatase, and the different treatments of the screened core-hole potential have visible impact on pre-edge intensities and the peak ratio of the main edge; (4) there exist moderate differences in the relative edge alignment of the three codes with a standard deviation of about 0.2 eV, which arise from multiple contributions including the frozen core approximation, final state effects, and different approximations used for the self-energy correction. In conclusion, our benchmark study provides important standards for first-principles XAS simulations with broad impact in data-driven XAS analysis.

36 MATERIALS SCIENCE↗
DOE OSTI2022

Validation and parameterization of a novel physics-constrained neural dynamics model applied to turbulent fluid flow

We report, in fluid physics, data-driven models to enhance or accelerate time to solution are becoming increasingly popular for many application domains, such as alternatives to turbulence closures, system surrogates, or for new physics discovery. In the context of reduced order models of high-dimensional time-dependent fluid systems, machine learning methods grant the benefit of automated learning from data, but the burden of a model lies on its reduced-order representation of both the fluid state and physical dynamics. In this work, we build a physics-constrained, data-driven reduced order model for Navier–Stokes equations to approximate spatiotemporal fluid dynamics in the canonical case of isotropic turbulence in a triply periodic box. The model design choices mimic numerical and physical constraints by, for example, implicitly enforcing the incompressibility constraint and utilizing continuous neural ordinary differential equations for tracking the evolution of the governing differential equation. We demonstrate this technique on a three-dimensional, moderate Reynolds number turbulent fluid flow. In assessing the statistical quality and characteristics of the machine-learned model through rigorous diagnostic tests, we find that our model is capable of reconstructing the dynamics of the flow over large integral timescales, favoring accuracy at the larger length scales. More significantly, comprehensive diagnostics suggest that physically interpretable model parameters, corresponding to the representations of the fluid state and dynamics, have attributable and quantifiable impact on the quality of the model predictions and computational complexity.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2025

Multi-Omics Reveals Temporal Scales of Carbon Metabolism in Synechococcus Elongatus PCC 7942 Under Light Disturbance

Central carbon metabolism in model cyanobacteria involves multiple pathways to adapt to energy-light limitations across diel cycles. However, the success in mechanistic modeling for phenotypic prediction of the protein regulators in the metabolic state depends on capturing the vast possibilities emerging from multiple regulatory pathways in complex biological processes. Here, we developed a physics-informed machine learning approach based on energy-landscape concepts to predict regulatory proteins responding to cyclic circadian and unforeseen light perturbations in cyanobacterial metabolic networks. Our approach provides interpretable de novo models for inferring gene expression dynamics from Synechococcus elongatus over diel cycles and using redox proteome analysis to distinguish immediate light-responsive elements from circadian-regulated processes in carbon metabolism pathways. We identified distinct temporal signatures with the analysis of the redox proteome: there was an immediate shift in cysteine redox states accompanied by a limited change in protein abundance under constant illumination and after 2 hours of darkness. This discovery indicates that the generation of reductants coordinates photoinduced electron transport with redox metabolic pathways in two discernable molecular mechanisms: fast redox-based protein modifications occur immediately after the light disturbance, followed by slow transcriptional regulations across networks. This temporal regulation reveals how metabolic networks integrate rapid light responses with programmed circadian rhythms to maintain cellular homeostasis under the light-energy limitations over the diel cycle.

Biomolecular & subcellular processes↗
DOE OSTI2025

Feature engineering descriptors, transforms, and machine learning for grain boundaries and variable-sized atom clusters

Abstract Obtaining microscopic structure-property relationships for grain boundaries is challenging due to their complex atomic structures. Recent efforts use machine learning to derive these relationships, but the way the atomic grain boundary structure is represented can have a significant impact on the predictions. Key steps for property prediction common to grain boundaries and other variable-sized atom clustered structures include: (1) describing the atomic structure as a feature matrix, (2) transforming the variable-sized feature matrix to a fixed length common to all structures, and (3) applying a machine learning algorithm to predict properties from the transformed matrices. We examine how these steps and different combinations of engineered features impact the accuracy of grain boundary energy predictions using a database of over 7000 grain boundaries. Additionally, we assess how different engineered features support interpretability, offering insights into the physics of the structure-property relationships.

36 MATERIALS SCIENCE↗
DOE OSTI2017

Discovering phases, phase transitions, and crossovers through unsupervised machine learning: A critical examination

Here, we apply unsupervised machine learning techniques, mainly principal component analysis (PCA), to compare and contrast the phase behavior and phase transitions in several classical spin models - the square and triangular-lattice Ising models, the Blume-Capel model, a highly degenerate biquadratic-exchange spin-one Ising (BSI) model, and the 2D XY model, and examine critically what machine learning is teaching us. We find that quantified principal components from PCA not only allow exploration of different phases and symmetry-breaking, but can distinguish phase transition types and locate critical points. We show that the corresponding weight vectors have a clear physical interpretation, which is particularly interesting in the frustrated models such as the triangular antiferromagnet, where they can point to incipient orders. Unlike the other well-studied models, the properties of the BSI model are less well known. Using both PCA and conventional Monte Carlo analysis, we demonstrate that the BSI model shows an absence of phase transition and macroscopic ground-state degeneracy. The failure to capture the 'charge' correlations (vorticity) in the BSI model (XY model) from raw spin configurations points to some of the limitations of PCA. Finally, we employ a nonlinear unsupervised machine learning procedure, the 'antoencoder method', and demonstrate that it too can be trained to capture phase transitions and critical points.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2024

How the Galaxy–Halo Connection Depends on Large-scale Environment

We investigate the connection between galaxies, dark matter halos, and their large-scale environments at z = 0 with Illustris TNG300 hydrodynamic simulation data. We predict stellar masses from subhalo properties to test two types of machine learning (ML) models: explainable boosting machines (EBMs) with simple galaxy environment features and E(3)-invariant graph neural networks (GNNs). The best-performing EBM models leverage spherically averaged overdensity features on 3 Mpc scales. Interpretations via SHapley Additive exPlanations also suggest that in the context of the TNG300 galaxy–halo connection, simple spherical overdensity on ∼3 Mpc scales is more important than cosmic web distance features measured using the DisPerSE algorithm. Meanwhile, a GNN with connectivity defined by a fixed linking length, L, outperforms the EBM models by a significant margin. As we increase the linking length scale, GNNs learn important environmental contributions up to the largest scales we probe (L = 10 Mpc). We conclude that 3 Mpc distance scales are most critical for describing the TNG galaxy–halo connection using the spherical overdensity parameterization, but that information on larger scales, which is not captured by simple environmental parameters or cosmic web features, can further augment these models. Our study highlights the benefits of using interpretable ML algorithms to explain models of astrophysical phenomena, and the power of using GNNs to flexibly learn complex relationships directly from data while imposing constraints from physical symmetries.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2021

Interpretable artificial intelligence and exascale molecular dynamics simulations to reveal kinetics: Applications to Alzheimer's disease

The rapid increase in computing power, especially with the integration of graphics processing units, has dramatically increased the capabilities of molecular dynamics simulations. To date, these capabilities extend from running very long simulations (tens to hundreds of microseconds) to thousands of short simulations. However, the expansive data generated in these simulations must be made interpretable not only by the investigator who performs them but also by others as well. Here, we demonstrate how integrating learning techniques, such as artificial intelligence, machine learning, and neural networks, into analysis pipelines can reveal the kinetics of Alzheimer's disease (AD) protein aggregation. Finally, we review select AD targets, describe current simulation methods, and introduce learning concepts and their application in AD, highlighting limitations and potential solutions.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2020

Automating the interpretation of PM 2.5 time–resolved measurements using a data–driven approach

The rapid development of automated measurement equipment enables researchers to collect greater quantities of time-resolved data from indoor and outdoor environments. While significant, the interpretation of the resulting data can be a time-consuming effort. This paper introduces an automated process of interpreting PM 2.5 time-resolved data and differentiating PM 2.5 emissions resulting from indoor and outdoor sources. Here, we use Random Forest (RF), a machine learning approach, to study a dataset of 836 indoor emission events that occurred over a 2-week period in 18 apartments in California. In this paper, we show model development and evaluate its performance as the sample size and source vary. We discuss the characteristics of the dataset that tended to help the source identification and why. For example, we show that data from many events and from different apartments are essential for the model to be suitable for analyzing a new separate dataset. We also show that longitudinal data appear to be more helpful than the time frequency of measurements within a given apartment. We use the resulting RF model to analyze PM 2.5 data of an entirely separate dataset collected from 65 new homes in California. The RF model identifies 442 indoor emission events, with only a few misidentifications.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2026

Wolf

The Workflow Orchestration Language Framework (WOLF) is an agentic framework grounded in natural language with an architecture inspired by reinforcement learning (RL)—designed to orchestrate, scale, and accelerate complex workflows. The concept of WOLF was born out of the very successful ASC Tri-lab Multi-Agent Design Assistant (MADA) project, but extends beyond its domain-specific design agents to provide a more general and extensible architecture. WOLF capitalizes on the lessons learned from MADA and is fully aligned with Sutton’s The Bitter Lesson—that the most enduring progress in AI comes from general-purpose methods that scale with computation, rather than narrow techniques built on domain-specific human knowledge. In this spirit, WOLF enables agents to autonomously learn workflows, capture strategies as reusable playbooks, and build a growing corpus of interpretable, auditable “wisdom artifacts.” These artifacts, expressed in natural language, bridge human and machine understanding while preserving adaptability and scalability as computational power continues to expand.

Boureima, Ismaeal↗
DOE OSTI2022

Geometric learning for computational mechanics Part II: Graph embedding for interpretable multiscale plasticity

The history-dependent behaviors of classical plasticity models are often driven by internal variables evolved according to phenomenological laws. The difficulty to interpret how these internal variables represent a history of deformation, the lack of direct measurement of these internal variables for calibration and validation, and the weak physical underpinning of those phenomenological laws have long been criticized as barriers to creating realistic models. In this work, geometric machine learning on graph data (e.g. finite element solutions) is used as a means to establish a connection between nonlinear dimensional reduction techniques and plasticity models. Geometric learning-based encoding on graphs allows the embedding of rich time-history data onto a low-dimensional Euclidean space such that the evolution of plastic deformation can be predicted in the embedded feature space. Finally, a corresponding decoder can then convert these low-dimensional internal variables back into a weighted graph such that the dominating topological features of plastic deformation can be observed and analyzed.

42 ENGINEERING↗
DOE OSTI2019

The Role of Chemistry in Fracture Pattern Development and Opportunities to Advance Interpretations of Geological Materials

Abstract Fracture pattern development has been a challenging area of research in the Earth sciences for more than 100 years. Much has been learned about the spatial and temporal complexity inherent to these systems, but severe challenges remain. Future advances will require new approaches. Chemical processes play a larger role in opening‐mode fracture pattern development than has hitherto been appreciated. This review examines relationships between mechanical and geochemical processes that influence the fracture patterns recorded in natural settings. For fractures formed in diagenetic settings (~50 to 200 °C), we review evidence of chemical reactions in fractures and show how a chemical perspective helps solve problems in fracture analysis. We also outline impediments to subsurface pattern measurement and interpretation, assess implications of discoveries in fracture history reconstruction for process‐based models, review models of fracture cementation and chemically assisted fracture growth, and discuss promising paths for future work. To accurately predict the mechanical and fluid flow properties of fracture systems, a processes‐based approach is needed. Progress is possible using observational, experimental, and modeling approaches that view fracture patterns and properties as the result of coupled mechanical and chemical processes. A critical area is reconstructing patterns through time. Such data sets are essential for developing and testing predictive models. Other topics that need work include models of crystal growth and dissolution rates under geological conditions, cement mechanical effects, and subcritical crack propagation. Advances in machine learning and 3‐D imaging present opportunities for a mechanistic understanding of fracture formation and development, enabling prediction of spatial and temporal complexity over geologic timescales. Geophysical research with a chemical perspective is needed to correctly identify and interpret fractures from geophysical measurements during site characterization and monitoring of subsurface engineering activities.

58 GEOSCIENCES↗
DOE OSTI2026

Feedback, physics, and forecasts: The emerging paradigm of machine learning-driven battery research

Machine learning (ML) is reshaping how we understand, predict, and optimize electrochemical systems. In batteries, ML accelerates discovery across chemistry, design, and operation by transforming massive experimental and simulated datasets into predictive, interpretable models. This review consolidates a decade of progress in ML-driven battery innovation, from early-cycle feature extraction to operando image analysis and physics-informed modeling. We categorize approaches by data domain and physical fidelity, emphasizing interpretable ML for diagnostics, reinforcement learning for control, and multi-objective optimization for lifetime extension strategies. Additionally, we demonstrate how integrated models accelerate discovery, reduce testing time, and guide sustainable design. Economic analyses furthermore illustrate how these advances can lower cost per cycle and improve circularity. Together, these developments chart a path toward self-optimizing, sustainable battery technologies.

artificial intelligence↗
DOE OSTI2025

A structured framework for predicting sustainable aviation fuel properties using liquid-phase FTIR and machine learning

Sustainable aviation fuels have the potential to improve efficiency, reduce emissions, and enhance energy security. To help identify viable sustainable aviation fuels and accelerate research, machine learning models have been developed to predict relevant physicochemical properties. However, many models have limited applicability, leverage data from complex analytical techniques with confined spectral ranges, or use feature decomposition methods that offer limited interpretability. Using liquid-phase Fourier Transform Infrared (FTIR) spectra, this study presents a structured method for creating accurate and interpretable property prediction models for neat molecules, aviation fuels, and blends. Liquid FTIR spectra can be collected quickly and consistently, offering high reliability, sensitivity, and component specificity using less than 2 ml of sample. The method first decomposes FTIR spectra into fundamental building blocks using non-negative matrix factorization (NMF) to enable scientific analysis of FTIR spectra attributes and fuel properties. The NMF features are then used to create five ensemble models for predicting final boiling point, flash point, freezing point, density at 15°C, and kinematic viscosity at -20°C. All models were trained using experimental property data from neat molecules, aviation fuels, and blends. The models accurately predict key properties across a broad range of neat molecules and representative fuels and blends, while enabling interpretation of relationships between compositional elements, such as functional groups or chemical classes, and their resulting properties. This demonstrates strong potential to support sustainable aviation fuel research and development. The models and data are available on an interactive web tool.

Fourier transform infrared spectroscopy↗
DOE OSTI2021

Explore Spatio‐Temporal Learning of Large Sample Hydrology Using Graph Neural Networks

Abstract Streamflow forecasting over gauged and ungauged basins play a vital role in water resources planning, especially under the changing climate. Increased availability of large sample hydrology data sets, together with recent advances in deep learning techniques, has presented new opportunities to explore temporal and spatial patterns in hydrological signatures for improving streamflow forecasting. The purpose of this study is to adapt and benchmark several state‐of‐the‐art graph neural network (GNN) architectures, including ChebNet, Graph Convolutional Network (GCN), and GraphWaveNet, for end‐to‐end graph learning. We explicitly represent river basins as nodes in a graph, learn the spatiotemporal nodal dependencies, and then use the learned relations to predict streamflow simultaneously across all nodes in the graph. The efficacy of the developed GNN models is investigated using the Catchment Attributes and MEteorology for Large‐sample Studies (CAMELS) data set under two settings, fixed graph topology (transductive learning), and variable graph topology (inductive learning), with the latter applicable to prediction in ungauged basins (PUB). Results indicate that GNNs are generally robust and computationally efficient, achieving similar or better performance than a baseline model trained using the long short‐term memory (LSTM) network. Further analyses are conducted to interpret the graph learning process at the edge and node levels and to investigate the effect of different model configurations. We conclude that graph learning constitutes a viable machine learning‐based method for aggregating spatiotemporal information from a multitude of sources for streamflow forecasting

Sun, Alexander Y.↗
DOE OSTI2019

An application of survival analysis to disruption prediction via Random Forests

One of the most pressing challenges facing the fusion community is adequately mitigating or, even better, avoiding disruptions of tokamak plasmas. However, before this can be done, disruptions must first be predicted with sufficient warning time to actuate a response. The established field of survival analysis provides a convenient statistical framework for time-to-event (i.e. time-to-disruption) studies. This paper demonstrates the integration of an existing disruption prediction machine learning algorithm with the Kaplan-Meier estimator of survival probability. Specifically discussed are the implied warning times from binary classification of disruption databases and the interpretation of output signals from Random Forest algorithms trained and tested on these databases. This survival analysis approach is applied to both smooth and noisy test data to highlight important features of the survival and hazard functions. In addition, this method is applied to three Alcator C-Mod plasma discharges and compared to a threshold-based scheme for triggering alarms. In one case, both techniques successfully predict the disruption; although, in another, neither warns of the impending disruption with enough time to mitigate. For the final discharge, the survival analysis approach could avoid the false alarm triggered by the threshold method. Limitations of this analysis and opportunities for future work are also presented.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2022

Nanoscale defect evaluation framework combining real-time transmission electron microscopy and integrated machine learning-particle filter estimation

Observation of dynamic processes by transmission electron microscopy (TEM) is an attractive technique to experimentally analyze materials’ nanoscale phenomena and understand the microstructure-properties relationships in nanoscale. Even if spatial and temporal resolutions of real-time TEM increase significantly, it is still difficult to say that the researchers quantitatively evaluate the dynamic behavior of defects. Images in TEM video are a two-dimensional projection of three-dimensional space phenomena, thus missing information must be existed that makes image’s uniquely accurate interpretation challenging. Therefore, even though they are still a clustering high-dimensional data and can be compressed to two-dimensional, conventional statistical methods for analyzing images may not be powerful enough to track nanoscale behavior by removing various artifacts associated with experiment; and automated and unbiased processing tools for such big-data are becoming mission-critical to discover knowledge about unforeseen behavior. We have developed a method to quantitative image analysis framework to resolve these problems, in which machine learning and particle filter estimation are uniquely combined. The quantitative and automated measurement of the dislocation velocity in an Fe-31Mn-3Al-3Si autunitic steel subjected to the tensile deformation was performed to validate the framework, and an intermittent motion of the dislocations was quantitatively analyzed. The framework is successfully classifying, identifying and tracking nanoscale objects; these are not able to be accurately implemented by the conventional mean-path based analysis.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Leveraging interpolation models and error bounds for verifiable scientific machine learning

Effective verification and validation techniques for modern scientific machine learning workflows are challenging to devise. Statistical methods are abundant and easily deployed, but often rely on speculative assumptions about the data and methods involved. Error bounds for classical interpolation techniques can provide mathematically rigorous estimates of accuracy, but often are difficult or impractical to determine computationally. Here, in this work, we present a best-of-both-worlds approach to verifiable scientific machine learning by demonstrating that (1) multiple standard interpolation techniques have informative error bounds that can be computed or estimated efficiently; (2) comparative performance among distinct interpolants can aid in validation goals; (3) deploying interpolation methods on latent spaces generated by deep learning techniques enables some interpretability for black-box models. We present a detailed case study of our approach for predicting lift-drag ratios from airfoil images. Code developed for this work is available in a public Github repository.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

State predictive information bottleneck

We report the ability to make sense of the massive amounts of high-dimensional data generated from molecular dynamics simulations is heavily dependent on the knowledge of a low-dimensional manifold (parameterized by a reaction coordinate or RC) that typically distinguishes between relevant metastable states, and which captures the relevant slow dynamics of interest. Methods based on machine learning and artificial intelligence have been proposed over the years to deal with learning such low-dimensional manifolds, but they are often criticized for a disconnect from more traditional and physically interpretable approaches. To deal with such concerns, in this work we propose a deep learning based state predictive information bottleneck approach to learn the RC from high-dimensional molecular simulation trajectories. We demonstrate analytically and numerically how the RC learnt in this approach is connected to the committor in chemical physics and can be used to accurately identify transition states. A crucial hyperparameter in this approach is the time delay or how far into the future the algorithm should make predictions about. Through careful comparisons for benchmark systems, we demonstrate that this hyperparameter choice gives useful control over how coarse-grained we want the metastable state classification of the system to be. We thus believe that this work represents a step forward in systematic application of deep learning based ideas to molecular simulations.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗