SEARCH · Engineering Papers

Results for “Generative Neural Networks”

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 667 records · Page 37

All sources NASA NTRS DOE OSTI

DOE OSTI2026

RU Net for Automatic Characterization of TRISO Fuel Cross Sections

TRistructural ISOtropic (TRISO) particle fuel is a type of nuclear fuel known for its high-temperature and high-burnup performance. Each sub-millimeter diameter TRISO particle consists of uranium-oxycarbide (UCO) or UO2 fuel kernel, coated with buffer, inner pyrolytic carbon (IPyC), silicon carbide (SiC), and outer pyrolytic carbon (OPyC) layers. The SiC layer acts as the main containment barrier for the TRISO particle to retain the fission products, while the IPyC and OPyC layers provide additional barriers to the release of fission products, especially fission gases. During irradiation, phenomena like kernel swelling, buffer densification, and IPyC fracture may impact fuel performance. Post-irradiation microscopy on entire compact cross sections or samples of individual particles deconsolidated from compacts is often used to identify these irradiation-induced changes in morphology. However, each fuel compact generally contains thousands of TRISO particles. To get statistical information on these phenomena, it is cumbersome work if done manually. For example, to get information about swelling/densification behaviors of different layers or kernels after irradiation, researchers previously manually measured the perimeter of each TRISO layer in hundreds of particles after four rounds of iterative grinding and polishing encompassing more than 2000 cross-section images for a total of four fuel compacts. To attempt to reduce the subjectivity inherent in that process and accelerate data analysis, we conducted a study on the automatic TRISO layer segmentation on cross-sectional microscopic images using Convolutional Neural Networks (CNNs). CNNs are a class of machine learning algorithms specifically designed for processing structured grid data that have gained popularity in recent years due to their remarkable performance in various computer vision tasks, including image classification, object detection, and image segmentation. In this research, we have generated the large irradiated TRISO layer dataset with more than 2000 cross-section TRISO microscopic images and the corresponding annotated images. Based on these annotated images, we have employed different CNNs for automatic segmentation of different TRISO layers. These include RU-Net (developed in this study), as well as three existing architectures: U-Net, Residual Network (ResNet), and Attention U-Net. The preliminary results show that the model based on RU-Net has the best performance in terms of intersection-over-union (IoU). Through the aid of these CNN models, we can expedite the analysis of TRISO particle cross-sections, significantly reducing the manual labor involved and improving the objectivity of the segmentation results.

Convolutional Neural Networks↗

DOE OSTI2024

Search for Long-Lived Heavy Neutral Leptons with Lepton Flavour Conserving or Violating Decays to a Jet and a Charged Lepton

A search for long-lived heavy neutral leptons (HNLs) is presented, which considers the hadronic final state and coupling scenarios involving all three lepton generations in the 2–20 GeV HNL mass range for the first time. Events comprising two leptons (electrons or muons) and jets are analyzed in a data sample of proton-proton collisions, recorded with the CMS experiment at the CERN LHC at a centre-of-mass energy of 13 TeV, corresponding to an integrated luminosity of 138 fb$^{−1}$. A novel jet tagger, based on a deep neural network, has been developed to identify jets from an HNL decay using various features of the jet and its constituent particles. The network output can be used as a powerful discriminating tool to probe a broad range of HNL lifetimes and masses. Contributions from background processes are determined from data. No excess of events in data over the expected background is observed. Upper limits on the HNL production cross section are derived as functions of the HNL mass and the three coupling strengths V$_{ℓN}$ to each lepton generation ℓ and presented as exclusion limits in the coupling-mass plane, as lower limits on the HNL lifetime, and on the HNL mass. In this search, the most stringent limit on the coupling strength is obtained for pure muon coupling scenarios; values of |$ {V}_{\mu \textrm{N}}^2 $| > 5 (4) × 10$^{−7}$ are excluded for Dirac (Majorana) HNLs with a mass of 10 GeV at a confidence level of 95% that correspond to proper decay lengths of 17 (10) mm.[graphic not available: see fulltext]

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

DOE OSTI2020

Learning Optimal Solutions for Extremely Fast AC Optimal Power Flow

We develop, in this paper, a machine learning approach to optimize the real-time operation of electric power grids. In particular, we learn feasible solutions to the AC optimal power flow (OPF) problem with negligible optimality gaps. The AC OPF problem aims at identifying optimal operational conditions of the power grids that minimize power losses and/or generation costs. Due to the computational challenges with solving this nonconvex problem, many efforts have focused on linearizing or approximating the problem in order to solve the AC OPF on faster timescales. However, many of these approximations can be fairly poor representations of the actual system state and still require solving an optimization problem, which can be time consuming for large networks. In this work, we learn a mapping between the system loading and optimal generation values, enabling us to find near-optimal and feasible AC OPF solutions. This allows us to bypass solving the traditionally nonconvex AC OPF problem, resulting in a significant decrease in computational burden for grid operators.

machine learning↗

DOE OSTI2020

Learning Optimal Solutions for Extremely Fast AC Optimal Power Flow: Preprint

machine learning↗

DOE OSTI2025

Toward accelerating rare-earth metal extraction using equivariant neural networks

The separation of rare-earth metals, vital for numerous advanced technologies, is hampered by their similar chemical properties, making ligand discovery a significant challenge. Traditional experimental and quantum chemistry approaches for identifying effective ligands are often resource-intensive. We introduce a machine learning protocol based on an equivariant neural network, Allegro, for the rapid and accurate prediction of binding energies in rare-earth complexes. Key to this work is our newly curated dataset of rare-earth metal complexes—made publicly available to foster further research—systematically generated using the Architector program. This dataset distinctively features functionalized derivatives of proven rare-earth-chelating scaffolds, hydroxypyridinone (HOPO), catecholamide (CAM), and their thio-analogues, selected for their established efficacy in binding these elements. Trained on this valuable resource, our Allegro models demonstrate excellent performance, particularly when trained to directly predict DFT-level binding energies, yielding highly accurate results that closely correlate with theoretical calculations on a diverse test set. Furthermore, this strategy exhibited strong out-of-sample generalization, accurately predicting binding energies for an isomeric HOPO-derivative ligand not seen during training. By substantially reducing computational demands, this machine learning framework, alongside the provided dataset, represent powerful tools to accelerate the high-throughput screening and rational design of novel ligands for efficient rare-earth metal separation.

Gupta, Ankur K. [Lawrence Berkeley National Labora↗

DOE OSTI2025

Machine learning-enabled discovery of ionic liquid–solvent electrolytes exhibiting high ionic conductivity

Ionic liquids (ILs), which are a class of materials with versatile nature and growing popularity, are facing impediments toward widespread usage as electrolytes due to various factors such as low ionic conductivity, high viscosity, high market price etc. One of the ways these limitations can be addressed is by mixing ILs with a molecular solvent. In a combinatorial sense, there exists an immense number of specific IL–solvent combinations. An exhaustive experimental or even simulation-based investigation of the chemical space spanned by such combinations can be extremely time-consuming, expensive, and nearly impossible. An alternative approach is to employ machine learning-based models developed from available databases. Although there exists prior literature that integrates machine learning to investigate mixtures of specific solvents with ILs, these models lack generalization necessitating development of a large number of ML models to handle various solvents. To remedy this shortcoming, as a part of designing green electrolytes with high ionic conductivity that can have potential applications in next-generation batteries and solar cells, this work aims to develop a unified machine learning model to predict ionic conductivity of any IL–solvent mixture system. In this regard, three models, namely, Random Forest, extreme gradient boosting (XGBoost), and artificial neural network (ANN) were formulated using the NIST ILThermo database. The dataset contained 549 unique ionic liquids from 16 cation families and 81 unique solvents, representing a total of 23 712 datapoints. SHAPLEY additive explanation (SHAP) method was used to assess the impact of various features on model prediction and their significance was compared with literature to gain physical insight about the model behavior. Finally, using the developed models, approximately 2.5 million IL–solvent mixtures at five different compositions were screened at room temperature. The high-throughput screening yielded nearly 19 000 IL–solvent mixtures for which ionic conductivity was found to exceed the ionic conductivity of conventional Li-ion battery electrolyte.

25 ENERGY STORAGE↗

DOE OSTI2020

Deep learning surrogate model for kinetic Landau-fluid closure with collision

In this work, the kinetic Landau-fluid (LF) closure with collision and periodic boundary condition is used in the development of the deep learning (DL) surrogate model. A classical neural network, namely, feedforward neural network or sometimes termed multilayer perceptron, is constructed and trained to learn the kinetic LF closure in the static limit and arbitrary mean free path in configuration space. The preliminary relation between best hyperparameters and critical parameters for data generation is found. Compared with the numerical approach (non-Fourier method) of the LF closure, the deep learning surrogate model shows an order of magnitude of improvement in terms of accuracy. Perhaps most importantly, the surrogate model closure has been integrated for the first time with fluid simulations. Our DL-enabled fluid simulations, for the first time, give the correct Landau damping rate for a wide range of wave vectors, while the Hammett–Perkins closure cannot produce the correct damping rate. We correctly connect the collisionless Hammett–Perkins closure and collisional Braginskii closure to reproduce the intrinsic nonlocal feature of the heat flux with DL techniques. We address the most concerning error accumulation problem and find that simulations with the deep learning surrogate model are as good as, if not better than, simulations with the analytic closure in terms of long-term numerical stability in the linear Landau damping test.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

DOE OSTI2025

GRUMDN: A Multi-Task Model for Predicting Human Patterns-of-Life from Stay Transition Data

Understanding human patterns-of-life (PoL) is essential towards ensuring safe and secure indoor facility environment as well as outdoor urban environment. Prediction of human movement in between places of interest is vital in understanding human PoL. Movement between spaces maybe represented and detected in one of the two forms: 1) trajectories: locations measured at regular time intervals by mobile sensors, bluetooth or GPS sensors; or 2) stay transitions: semantic PoI (points of interest) and stay duration data measurable by eventbased sensors that collect data when a check-in or check-out event is detected. Stay transition data provides a more compressed data format compared to trajectories data, especially in situations with longer stay durations, while preserving the information necessary for PoL analysis. Now as introduced briefly in the paper, our deployed end application (Digital Twin of a facility with non-player characters, besides the interactive user in virtual reality) needed a well-performing and validated AI/ML model for simulating high quality stay transitions behavior. In this study we thus primarily present our findings with developing and validating that model, which is a multi-task neural network for stay transition prediction. The neural network consists of two heads, for corresponding two tasks of stay category prediction and stay duration prediction. We evaluated gated recurrent units and multi-layer perceptrons of varying network sizes for stay category prediction; while mixture density networks, noisy generator-only networks, and generative adversarial networks of varying network sizes for stay duration prediction. We have then evaluated four multi-task models, constructed by combining these specialized models, on their ability to predict stay transition data. We tested our models on datasets from two different cases: 1) a simulation-generated dataset of indoor movement within the HFIR (high flux isotope reactor) nuclear reactor facility at Oak Ridge National Laboratory (ORNL); and 2) the GeoLife human mobility dataset of outdoor urban movement available in literature. Our results indicate that GRUMDN, which combines gated recurrent units (GRU) for stay category prediction task, and mixture density networks (MDN) for stay duration prediction task, did overall outperform other multitask models and the current state-of-the-art.

Gunaratne, Chathika [ORNL] (ORCID:0000000225088745↗

DOE OSTI2024

Online MCMC Thinning with Kernelized Stein Discrepancy

A fundamental challenge in Bayesian inference is efficient representation of a target distribution. Many nonparametric approaches do so by sampling a large number of points using variants of Markov chain Monte Carlo (MCMC). Here, we propose an MCMC variant that retains only those posterior samples which exceed a kernelized Stein discrepancy (KSD) threshold, which we call KSD thinning. We establish the convergence and complexity trade-offs for several settings of KSD thinning as a function of the KSD threshold parameter, sample size, and other problem parameters. We provide experimental comparisons against other online nonparametric Bayesian methods that generate low-complexity posterior representations. We observe superior consistency/complexity trade-offs across a range of settings including MCMC sampling on two Bayesian inference problems from the biological sciences, and 10 × inference speedup and storage reduction for Bayesian neural networks with no loss of accuracy and no increase in training time. Our code is available at https://github.com/colehawkins/KSD-Thinning.

Bayesian inference↗

DOE OSTI2026

Iterative ML and Experiments for Emerging VOCs

SAND2026-17074O Iterative ML and Experiments for Emerging VOCs is a tool that analyzes and predicts the behaviors of SARS-CoV-2 variants. It processes experimental data on ACE2 (the receptor for the SARS-CoV-2 virus that allows it to infect the cell) and antibody binding using machine learning models, including neural networks, to forecast ACE2 interactions and variant expression. The tool employs transfer learning and global epistasis modeling, integrating public datasets with proprietary data to enhance prediction accuracy. Additionally, it fits concentration-response curves to determine dissociation constants and generates visualizations to support research findings, thereby aiding in the identification of new antibodies for emerging variants of concern. Sandia National Laboratories is a multimission laboratory managed and operated by National Technology & Engineering Solutions of Sandia, LLC, a wholly owned subsidiary of Honeywell International Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA0003525.

Sheffield, Thomas [Sandia National Lab. (SNL-NM), ↗

DOE OSTI2021

Learning curves for drug response prediction in cancer cell lines

Motivated by the size and availability of cell line drug sensitivity data, researchers have been developing machine learning (ML) models for predicting drug response to advance cancer treatment. As drug sensitivity studies continue generating drug response data, a common question is whether the generalization performance of existing prediction models can be further improved with more training data. We utilize empirical learning curves for evaluating and comparing the data scaling properties of two neural networks (NNs) and two gradient boosting decision tree (GBDT) models trained on four cell line drug screening datasets. The learning curves are accurately fitted to a power law model, providing a framework for assessing the data scaling behavior of these models. The curves demonstrate that no single model dominates in terms of prediction performance across all datasets and training sizes, thus suggesting that the actual shape of these curves depends on the unique pair of an ML model and a dataset. The multi-input NN (mNN), in which gene expressions of cancer cells and molecular drug descriptors are input into separate subnetworks, outperforms a single-input NN (sNN), where the cell and drug features are concatenated for the input layer. In contrast, a GBDT with hyperparameter tuning exhibits superior performance as compared with both NNs at the lower range of training set sizes for two of the tested datasets, whereas the mNN consistently performs better at the higher range of training sizes. Moreover, the trajectory of the curves suggests that increasing the sample size is expected to further improve prediction scores of both NNs. These observations demonstrate the benefit of using learning curves to evaluate prediction models, providing a broader perspective on the overall data scaling characteristics. A fitted power law learning curve provides a forward-looking metric for analyzing prediction performance and can serve as a co-design tool to guide experimental biologists and computational scientists in the design of future experiments in prospective research studies.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2020

Improving High-Energy Particle Detectors with Machine Learning

Microseconds after the Big Bang, the universe existed in a state called the quark-gluon plasma (QGP). To experimentally study its properties, the QGP is recreated in high-energy nuclear collisions at the LHC, and the particles produced from the QGP are reconstructed from their energy deposition in the ATLAS calorimeter. This requires both classifying the particles and calibrating their deposited energy. The objective of this project is to improve the reconstruction by using machine learning techniques, where the energy depositions of clusters of cells, formed by ATLAS topo-clustering methods, are treated as three-dimensional images when inputted to neural networks. This approach significantly improves the calibration of deposited energies when cross-validating while training, and models trained on idealized data predict the calibrated energies of particles in more complex data sets well. Additionally, implementation of a data generator using uproot allows the program to load input data into memory as needed while training or predicting, significantly reducing the amount of memory used. The data generator also allows for use of multiprocessing to speed up training and evaluating. This work illustrates that using machine learning methods for both classification and calibration has the potential to significantly improve particle reconstruction.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

DOE OSTI2023

Simulation Based Inference with Domain Adaptation for Strong Gravitational Lensing

Simulation based inference leverages machine learning to carry out Bayesian inference for systems with intractable likelihoods. However, transitioning a network trained on simulated data to real data runs the risk of encountering domain shift, leading to performance losses. We attempt to implement domain adaptation into the sbi neural posterior estimation framework using the Maximum Mean Discrepancy as an additional network loss. We test two network architectures and use masked autoregressive flow for density estimation. We test the network on a set of 400,000 simulated strong gravitational lensing images generated using deeplenstronomy. The source domain is defined as low noise whereas the target domain has a noise profile sampled from experimentally derived DES survey conditions. We find that, while DA does lead to performance improvements, they are marginal at ~6% less inference error. We also find a similar marginal improvement in uncertainty calibration at around 8%.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

DOE OSTI2025

Deep Point Cloud Building Envelope Segmentation (DeeP-CuBES) using Deep Learning

Building Information Modeling (BIM) plays an important role in building design and construction, particularly for achieving energy-efficient retrofits. Building envelope retrofits using panelized prefabricated system, such as those popularized by the Energiesprong program, need accurate as-built dimensions of facade features (windows, doors, etc.) to achieve the desired thermal and air tightness. Traditionally, building surveying is done manually, resulting in a time-consuming and labor-intensive process. Recently, 3D point clouds from terrestrial LiDAR have been used to automate the generation of as-built dimensions of existing buildings. However, automated BIM using LiDAR relies on solving the point cloud semantic segmentation (PCSS) problem. In this work, we propose a robust pipeline for solving the PCSS problem using deep neural networks, focusing on overcoming challenges posed by imbalanced datasets and complex architectural features. We introduce the first high-density, labeled, and validated building envelope point cloud dataset derived from multiple building scans, specifically curated to tackle challenges in facade-level segmentation. Results from the trained neural networks show that advanced attention-based architectures and incorporating radiometry (light intensity and RGB) features significantly boost segmentation accuracy for windows and doors.

Selvakumar, Balaji [ORNL]↗

DOE OSTI2022

Scaling Building Energy Audits through Machine Learning Methods on Novel Drone Image Data

Building energy audits are time-consuming and labor-intensive. This paper describes a new method using machine learning (ML) techniques on novel data sources (drone images) to improve the identification of building characteristics and retrofit opportunities, and thereby reduce the effort for audits. The new ML method includes: (1) Building footprint extraction using line extraction, polygonization, and polygon-merging, (2) Building envelope extraction using PIX4d modeling software to reconstruct a building 3D model, (3) Visualization tool for viewing images from the 3D model, (4) Window-to-wall ratio (WWR) using state-of-art deep neural network semantic segmentation, (5) Envelope thermal anomaly detection using an unsupervised machine learning clustering algorithm, and (6) Rooftop energy equipment detection based on an object detection algorithm. The testing of this method involved a comparison of additional ML-generated information overlaid on current ‘state-of-practice’ audit and remote assessment baselines using evaluation metrics: labor time and associated cost, marginal benefits of using ML-generated information in workflows for audits and remote assessments, integration potential with existing processes and tools, and replicability/scalability of the method. In two test buildings in California that had comprehensive drawings and meter data available, the ML method effectively generated a building footprint, envelope, rooftop equipment, WWR, and locations of envelope thermal anomalies. Projected target segments of the ML method are sites with minimal drawings and energy data, and underserved sectors such as multistoried housing, disadvantaged communities, and schools for which the ML method can enable identification of building asset characteristics and prioritization of envelope retrofits and decentralized energy equipment retrofits.

Singh, Reshma↗

DOE OSTI2025

Finding Real Uncertainties From Physical Simulations

Modeling strong gravitational lenses is computationally expensive for the complex data from modern and next-generation cosmic surveys. Deep learning has emerged as a promising approach for finding lenses and predicting lensing parameters, such as the Einstein radius. Mean-variance Estimators (MVEs) are a common approach for obtaining aleatoric (data) uncertainties from a neural network prediction. However, neural networks have not been demonstrated to perform well on out-of-domain target data successfully - e.g., when trained on simulated data and applied to real, observational data. In this work, we perform the first study of the efficacy of MVEs in combination with unsupervised domain adaptation (UDA) on strong lensing data. The source domain data is noiseless, and the target domain data has noise mimicking modern cosmology surveys. We find that adding UDA to MVE increases the accuracy on the target data by a factor of about two over an MVE model without UDA. Including UDA also permits much more well-calibrated aleatoric uncertainty predictions. Advancements in this approach may enable future applications of MVE models to real observational data.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

DOE OSTI2023

Results of the Photometric LSST Astronomical Time-series Classification Challenge (PLAsTiCC)

Abstract Next-generation surveys like the Legacy Survey of Space and Time (LSST) on the Vera C. Rubin Observatory (Rubin) will generate orders of magnitude more discoveries of transients and variable stars than previous surveys. To prepare for this data deluge, we developed the Photometric LSST Astronomical Time-series Classification Challenge (PLAsTiCC), a competition that aimed to catalyze the development of robust classifiers under LSST-like conditions of a nonrepresentative training set for a large photometric test set of imbalanced classes. Over 1000 teams participated in PLAsTiCC, which was hosted in the Kaggle data science competition platform between 2018 September 28 and 2018 December 17, ultimately identifying three winners in 2019 February. Participants produced classifiers employing a diverse set of machine-learning techniques including hybrid combinations and ensemble averages of a range of approaches, among them boosted decision trees, neural networks, and multilayer perceptrons. The strong performance of the top three classifiers on Type Ia supernovae and kilonovae represent a major improvement over the current state of the art within astronomy. This paper summarizes the most promising methods and evaluates their results in detail, highlighting future directions both for classifier development and simulation needs for a next-generation PLAsTiCC data set.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2024

DLSIA: Deep Learning for Scientific Image Analysis

DLSIA (Deep Learning for Scientific Image Analysis) is a Python-based machine learning library that empowers scientists and researchers across diverse scientific domains with a range of customizable convolutional neural network (CNN) architectures for a wide variety of tasks in image analysis to be used in downstream data processing. DLSIA features easy-to-use architectures, such as autoencoders, tunable U-Nets and parameter-lean mixed-scale dense networks (MSDNets). Additionally, this article introduces sparse mixed-scale networks (SMSNets), generated using random graphs, sparse connections and dilated convolutions connecting different length scales. For verification, several DLSIA-instantiated networks and training scripts are employed in multiple applications, including inpainting for X-ray scattering data using U-Nets and MSDNets, segmenting 3D fibers in X-ray tomographic reconstructions of concrete using an ensemble of SMSNets, and leveraging autoencoder latent spaces for data compression and clustering. As experimental data continue to grow in scale and complexity, DLSIA provides accessible CNN construction and abstracts CNN complexities, allowing scientists to tailor their machine learning approaches, accelerate discoveries, foster interdisciplinary collaboration and advance research in scientific image analysis.

97 MATHEMATICS AND COMPUTING↗