Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Automated labeling”

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 163 records · Page 9
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

Mapathons versus automated feature extraction: a comparative analysis for strengthening immunization microplanning

Background: Social instability and logistical factors like the displacement of vulnerable populations, the difficulty of accessing these populations, and the lack of geographic information for hard-to-reach areas continue to serve as barriers to global essential immunizations (EI). Microplanning, a population-based, healthcare intervention planning method has begun to leverage geographic information system (GIS) technology and geospatial methods to improve the remote identification and mapping of vulnerable populations to ensure inclusion in outreach and immunization services, when feasible. We compare two methods of accomplishing a remote inventory of building locations to assess their accuracy and similarity to currently employed microplan line-lists in the study area. Methods: The outputs of a crowd-sourced digitization effort, or mapathon, were compared to those of a machine-learning algorithm for digitization, referred to as automatic feature extraction (AFE). The following accuracy assessments were employed to determine the performance of each feature generation method: (1) an agreement analysis of the two methods assessed the occurrence of matches across the two outputs, where agreements were labeled as “befriended” and disagreements as “lonely”; (2) true and false positive percentages of each method were calculated in comparison to satellite imagery; (3) counts of features generated from both the mapathon and AFE were statistically compared to the number of features listed in the microplan line-list for the study area; and (4) population estimates for both feature generation method were determined for every structure identified assuming a total of three households per compound, with each household averaging two adults and 5 children. Results: The mapathon and AFE outputs detected 92,713 and 53,150 features, respectively. A higher proportion (30%) of AFE features were befriended compared with befriended mapathon points (28%). The AFE had a higher true positive rate (90.5%) of identifying structures than the mapathon (84.5%). The difference in the average number of features identified per area between the microplan and mapathon points was larger (t = 3.56) than the microplan and AFE (t = -2.09) (alpha = 0.05). Conclusions: Our findings indicate AFE outputs had higher agreement (i.e., befriended), slightly higher likelihood of correctly identifying a structure, and were more similar to the local microplan line-lists than the mapathon outputs. These findings suggest AFE may be more accurate for identifying structures in high-resolution satellite imagery than mapathons. However, they both had their advantages and the ideal method would utilize both methods in tandem.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2023

Machine Learning Automated Analysis of Enormous Synchrotron X-ray Diffraction Datasets

X-ray diffraction (XRD) data analysis can be a time-consuming and laborious task. Deep neural network (DNN) based models trained with synthetic XRD patterns have been proven to be a highly efficient, accurate, and automated method for analyzing common XRD data collected from solid samples in ambient environments. However, it remains unclear whether synthetic XRD-based models can be effective in solving micro(μ)-XRD mapping data for in situ experiments involving liquid phases, which always have lower quality and significant artifacts. In this study, we collected μ-XRD mapping data from a LaCl 3 -calcite hydrothermal fluid system and trained two categories of models to analyze the experimental XRD patterns. Here, the models trained solely with synthetic XRD patterns showed low accuracy (as low as 64%) when solving experimental μ-XRD mapping data. However, the accuracy of the DNN models significantly improved (90% or above) when we trained them with a data set containing both synthetic and a small number of labeled experimental μ-XRD patterns. This study highlights the importance of labeled experimental patterns in training DNN models to solve μ-XRD mapping data from in situ experiments involving liquid phases.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2020

Identifying Suspect Instrument Intervals Using Midnight Noise Time Histories

With an intense work ethic, and high levels of devotion to their craft, seismologists continue to battle the elements and wrestle with complex logistics in their efforts to field instrumentation over the entire globe. We are fortunate to see data volumes accelerating in size, and quality continuing to improve; for example, universal timing issues can now be considered rare. However, instrument response issues are still common, and can hinder research that seeks to understand amplitudes of seismic wavefields. We are interested in using the accumulations of seismic data to develop models that will predict high frequency (0.2-20 Hz, and higher) signal amplitudes for explosion monitoring purposes over broad areas, which requires large data sets and extensive quality control. To identify response and station health issues, we have collected noise time histories for global seismic data, focusing on measurements near (but not restricted to) midnight to eliminate diurnal variations, and have manually determined time intervals that appear inconsistent with background behavior. We assign descriptive labels, but do not attempt to diagnose causes. We use these intervals to discard data. To date, we have examined 39,260 channels from 11,105 stations, heavily weighted toward IRIS holdings, dates through 2017 (depending on station, roughly the date we started our manual review), bands between 1 and 8 Hz, finding 24,733 anomalous time intervals. The great majority (90%) of these intervals appear to be shifts of constant offset, often bounded by times of known instrument changes, likely the result of poor documentation of response parameters at one of many stages between the field and the plotter. We hope these results can be of use to our colleagues, and would encourage community efforts to diagnose anomalous behavior, and fix poor responses. We also hope these results will support automation efforts, including application of supervised learning techniques.

47 OTHER INSTRUMENTATION↗
DOE OSTI2025

Mapping Rare Earths and Toxics in E-Waste via Hyperspectral Imaging and Machine Learning

Electronic waste (e-waste) presents a mounting challenge to environmental sustainability due to its complex composition, which includes high-value rare earth elements, hazardous organic compounds, and non-recyclable plastics. Accurate and scalable material classification is essential for enabling efficient resource recovery and safe recycling practices. This study introduces a confidence-aware classification pipeline that combines mid-infrared hyperspectral imaging (HSI), spectral angle mapping (SAM), and iterative machine learning to perform pixel-level material identification across e-waste devices. A curated spectral library encompassing artificial materials (e.g., plastic iron oxide, galvanized metals), minerals (e.g., allanite, hematite), and organic compounds (e.g., benzanthracene, toluene) was used to generate pseudo-labels, each assigned a confidence score based on SAM-derived spectral similarity. High-confidence samples from seven consumer electronics—digital cameras, keyboards, laptop fans, modems, motherboards, TV remotes, and speakers—were iteratively expanded and classified using models such as Support Vector Machine (SVM), Random Forest, Gradient Boosting Classifier, Partial Least Squares Discriminant Analysis (PLSDA) and Logistic Regression. The best-performing classifiers achieved macro F1 scores approaching 1.0. Results revealed widespread plastic content (dominated by plastic iron oxide), the presence of rare earth-bearing minerals like cerium-containing allanite, and pervasive detection of hazardous organics such as benzanthracene. Principal Component Analysis (PCA) visualizations and confusion matrices confirmed high separability and robust classification performance. This methodology enables precise, non-destructive, and scalable classification of heterogeneous e-waste streams. It supports automated, hazard-aware sorting in recycling workflows, facilitating selective recovery of critical materials and compliance with circular economy goals. The confidence-aware framework provides a foundation for real-time deployment in industrial settings, offering significant implications for smart e-recycling infrastructure and policy-driven material stewardship.

Circular economy↗
DOE OSTI2024

FY24 Progress Report: SRNL Analysis of ICCWR LCM and WAMS data for Corrosion and Cracking

Algorithms for Machine Learning (ML) and data analysis for the 3013 Surveillance Program have been developed in an ongoing collaborative effort by the Savannah River National Laboratory (SRNL) and the University of South Carolina (USC). The objective of the algorithms is to automate the identification of corrosion and crack formation in the Inner Container Closure Weld Region (ICCWR) of the canister system used to store Pu-bearing material. Data for corrosion and cracking is collected from large binary files generated by a Laser Confocal Microscope (LCM), the Wide Area 3D Measurement System (WAMS), or,in a recent proposal, by a Scanning Electron Microscope (SEM). The ML software uses the physical attributes in the data files (e.g., one or more of: height, color, and 16-bit grayscale values as functions of position in a plane projection) to detect signs of surface corrosion and cracking after being trained on similar data, with the features to be detected. Although the initial scope included screening for broader indicators of corrosion, e.g., pitting, identification of potential cracks was prioritized for the past several years at the request of program leadership. Labeled training data is essential to developing the ML algorithm, and enhancements to data labeling capability have been developed to address this essential precursor to application of ML routines. Efficient labeling is particularly important in view of the large volume of data required to train ML algorithms and the relative rarity of cracks in the ICCWR data set. The updated program will read binary data from either LCM, WAMS or SEM files, interrogate data attributes, facilitate user labeling of data for training ML algorithms, execute ML algorithms, output parameters from trained ML algorithms, report ML model accuracy with respect to labeled data, and generate graphical representations for various analyses. In FY24, hourglass neural networks (HNNs) that were initiated in FY22 were further developed and tested using available LCM data, and their performance was tested against that of the alternative U-Net Neural Network algorithm structure. HNNs along with previously developed Convolutional Neural Networks (CNNs) and Deep Neural Networks (DNNs) comprise a suite of ML tools for identification of cracks in the ICCWR

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗
DOE OSTI2024

Evaluation of Digital Nautical Chart data for confirmation and expansion of GeoNames data

Here, this work examines how Digital Nautical Chart (DNC) data may contribute to the evolution and refinement of GeoNames data for near-shore features. GeoNames features are point data with one or more possible place names. DNC Earth Cover Text (ECRText) objects are map labels positioned nearby their real word counterpart. ECRText feature map position strikes a compromise between association with real features and cartographic readability. This work explores whether ECRText features can confirm (or expand names for) existing locations or contribute new locations through data conflation. Due to name variations and spatial position, conflating these data are nontrivial. Previous work engaged in a brief examination using the trigram string matching algorithm under coarse proximity constraints, indicating that ECRText could provide additional value to GeoNames. This work builds on that study, by engaging in a deeper examination of spatial proximity and exploring conflation agreement across an ensemble of string matching approaches. The result finds strong ensemble agreement about ECRText features which already exist in GeoNames but mixed results about which features contribute new information, as well as exploring why some of these matching techniques fail. With an eye toward automation, computational efficiency was found not to be a constraint in sustaining updates.

96 KNOWLEDGE MANAGEMENT AND PRESERVATION↗
DOE OSTI2025

Optimizing Facility Operations by Applying Machine Learning to the Army Reserve Enterprise Building Control System (Final Report)

Thousands of U.S. Department of Defense (DoD) buildings have building automation systems (BASs) and/or advanced meters. Although these systems have a wealth of data, performance optimization requires time and expertise to review and act on that information. Machine learning (ML) can provide automated and actionable insights to controls operators. This demonstration implemented proven ML methods on the Army Reserve Enterprise Building Control System. ML refers to algorithms that “learn” from data and improve their performance on a given task over time. In the buildings domain these tasks range from predicting future energy consumption, to identifying operational issues before faults occur, to optimizing control decisions. To learn, ML requires input data, which – for buildings – typically consists of instrument data such as energy consumption data and subsystem controls information such as set-point temperatures, and context data consisting of information such as the physical location of the building, the area of the building, and the weather. ML models use the relationships learned from the input data to make predictions with new, previously unseen, data. The team was able to investigate and successfully implement the following ML use cases: labeling consumption data as anomalous or non-anomalous; baseline whole-building load prediction (unknown fault status); fault detection (validation not possible); and site prioritization for energy-related projects. Due to the constraints of the project, interventions were not able to be implemented during the demonstration; therefore, assessments of operational cost savings and maintenance avoided could not be performed. The project has been presented at two leading national building conferences and two additional publications to peer-reviewed journals are currently in preparation.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2022

Semantic Property Graph for Scalable Knowledge Graph Analytics

Graphs are a natural and fundamental representation to describe entities, relationships, activities, and evolution of complex systems. Many domains such as communication, citation, procurement, biology, social media, and transportation can be modeled as a set of entities and their relationships. Resource Description Framework (RDF) and Labeled Property Graph (LPG) are two of the most used data models to encode information in a graph. Both models are similar in terms of using basic graph elements such as nodes and edges but differ in terms of the modeling approach, expressibility, serialization, and target applications. RDF is a flexible data exchange model for expressing information about entities but it tends to a have high memory footprint and inefficient storage, which does not make it a natural choice to perform scalable graph analytics. In contrast, LPG has gained traction as a reliable model to perform scalable graph analytic tasks such as sub-graph matching, network alignment, and real-time knowledge graph query. It provides efficient storage, fast traversal, and flexibility to model various real-world domains. At the same time, the LPG lacks the support of a formal knowledge representation such as an ontology to provide automated knowledge inference. We propose Semantic Property Graph (SPG) as a logical projection of reified RDF into the LPG model. SPG continues to use RDF ontology to define the type hierarchy of the projected graph and validate it against a given ontology. We present a framework to convert reified RDF graphs into SPG using two different computing environments. We also present cloud-based graph migration capabilities using Amazon Web Services.

Purohit, Sumit↗
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 OSTI2025

Machine learning model inputs, outputs, and scripts associated with “Artificial intelligence-guided iterations between observations and modeling significantly improve environmental predictions”

NOTE: The manuscript associated with this data package is currently in review. The data may be revised based on reviewer feedback. Upon manuscript acceptance, this data package will be updated with the final dataset and additional metadata. This data package is associated with the manuscript “Artificial intelligence-guided iterations between observations and modeling significantly improve environmental predictions” (Malhotra et al., in prep). This effort was designed following ICON (integrated, coordinated, open, and networked) principles to facilitate a model-experiment (ModEx) iteration approach, leveraging crowdsourced sampling across the contiguous United States (CONUS). New machine learning models were created every month to guide sampling locations. Data from the resulting samples were used to test and rebuild the machine learning models for the next round of sampling guidance. Associated sediment and water geochemistry and in situ sensor data can be found at https://data.ess-dive.lbl.gov/datasets/doi:10.15485/1923689, https://data.ess-dive.lbl.gov/datasets/doi:10.15485/1729719, and https://data.ess-dive.lbl.gov/datasets/doi:10.15485/1603775. This data package is associated with two GitHub repositories found at https://github.com/parallelworks/dynamic-learning-rivers and https://github.com/WHONDRS-Hub/ICON-ModEx_Open_Manuscript. In addition to this readme, this data package also includes two file-level metadata (FLMD) files that describes each file and two data dictionaries (DD) that describe all column/row headers and variable definitions. This data package consists of two main folders (1) dynamic-learning-rivers and (2) ICON-ModEx_Open_Manuscript which contain snapshots of the associated GitHub repositories. The input data, output data, and machine learning models used to guide sampling locations are within dynamic-learning-rivers. The folder is organized into five top-level directories: (1) “input_data” holds the training data for the ML models; (2) “ml_models” holds machine learning (ML) models trained on the data in “input_data”; (3) “examples” contains files for direct experimentation with the machine learning model, including scripts for setting up “hindcast” run; (4) “scripts” contains data preprocessing and postprocessing scripts and intermediate results specific to this data set that bookend the ML workflow; and (5) “output_data” holds the overall results of the ML model on that branch. Each trained ML model resides on its own branch in the repository; this means that inputs and outputs can be different branch-to-branch. There is also one hidden directory “.github/workflows”. This hidden directory contains information for how to run the ML workflow as an end-to-end automated GitHub Action but it is not needed for reusing the ML models archived here. Please see the top-level README.md in the GitHub repository for more details on the automation. The scripts and data used to create figures in the manuscript are within ICON-ModEx_Open_Manuscript. The folder is organized into four folders which contain the scripts, data, and pdf for each figure. Within the “fig-model-score-evolution” folder, there is a folder called “intermediate_branch_data” which contains some intermediate files pulled from dynamic-learning-rivers and reorganized to easily integrate into the workflows. NOTE: THIS FOLDER INCLUDES THE FILES AT THE POINT OF PAPER SUBMISSION. IT WILL BE UPDATED ONCE THE PAPER IS ACCEPTED WITH ANY REVISIONS AND WILL INCLUDE A DD/FLMD AT THAT POINT. We thank the United States Forest Service, Washington Department of Fish and Wildlife, Washington Department of Natural Resources, Cowiche Canyon Conservatory, Washington State Parks and Recreation Commission (Scientific Research Permit #210901), and the Confederated Tribes and Bands of the Yakama Nation for access to field locations where the samples labeled “SSS” were collected. We also thank the Yakama Nation Tribal Council and Yakama Nation Fisheries for working with us to facilitate sample collection and optimization of data usage according to their values and worldview. WHONDRS consortium members were asked to provide any acknowledgments for the collection of samples labeled “CM” and the following is a list of acknowledgments that were submitted with their corresponding Site IDs: (MART) Research activities were conducted in part on the Wind River Experimental Forest within the Gifford Pinchot National Forest; (MP- 100379) Philadelphia is part of Lenapehoking, the ancestral homelands of the Lenape peoples; (MP-102398) Land surveyed is the ancestral homelands of the Nookhose'iinenno (Arapaho), Tsis tsis'tas (Cheyenne), and Nuuchu (Ute); (MP-100749 and MP- 100747) Georgia Coastal Ecosystem LTER, OCE-1832178; (SP-70 and SP-72) Eastern Shoshone, Shoshone-Bannock; (MP- 102944) Funded by Oregon Watershed Enhancement Board. On the traditional lands of the Confederated Tribes of the Siletz, Confederated Tribes of the Grand Rhonde, and the Clatsop-Nehalem Confederated Tribe; (MP- 100607) Holiday Creek is located on the traditional territory of the Monacan Indian Nation; (SP-45) Lafayette Blue Springs State Park; (MP-102420) NSF DEB-2016749; (MP-100019) New Hampshire Agriculture Experiment Station; (SP-35) Rayonier (land owner; https://www.rayonier.com/); (MP- 101276) US Department of Energy, Office of Science, Biological and Environmental Research, Subsurface Biogeochemical Research, Watershed Dynamics and Evolution SFA at ORNL; (MP- 103224) Watershed Dynamics and Evolution SFA at ORNL; (MP- 101584) Traditional lands of the Oceti Sakowin (Dakota, Lakota, Nakoda) and Anishinaabe Peoples.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

Effect of image resolution on automated classification of chest X-rays

Deep learning (DL) models have received much attention lately for their ability to achieve expert-level performance on the accurate automated analysis of chest X-rays (CXRs). Recently available public CXR datasets include high resolution images, but state-of-the-art models are trained on reduced size images due to limitations on graphics processing unit memory and training time. As computing hardware continues to advance, it has become feasible to train deep convolutional neural networks on high-resolution images without sacrificing detail by downscaling. This study examines the effect of increased resolution on CXR classification performance. We used the publicly available MIMIC-CXR-JPG dataset, comprising 377,110 high resolution CXR images for this study. We applied image downscaling from native resolution to 2048 × 2048 pixels, 1024 × 1024 pixels, 512 × 512 pixels, and 256 × 256 pixels and then we used the DenseNet121 and EfficientNet-B4 DL models to evaluate clinical task performance using these four downscaled image resolutions. We find that while some clinical findings are more reliably labeled using high resolutions, many other findings are actually labeled better using downscaled inputs. We qualitatively verify that tasks requiring a large receptive field are better suited to downscaled low resolution input images, by inspecting effective receptive fields and class activation maps of trained models. Lastly, we show that stacking an ensemble across resolutions outperforms each individual learner at all input resolutions while providing interpretable scale weights, indicating that diverse information is extracted across resolutions.

47 OTHER INSTRUMENTATION↗
DOE OSTI2022

AICCA: AI-Driven Cloud Classification Atlas

Clouds play an important role in the Earth’s energy budget, and their behavior is one of the largest uncertainties in future climate projections. Satellite observations should help in understanding cloud responses, but decades and petabytes of multispectral cloud imagery have to date received only limited use. This study describes a new analysis approach that reduces the dimensionality of satellite cloud observations by grouping them via a novel automated, unsupervised cloud classification technique based on a convolutional autoencoder, an artificial intelligence (AI) method good at identifying patterns in spatial data. Our technique combines a rotation-invariant autoencoder and hierarchical agglomerative clustering to generate cloud clusters that capture meaningful distinctions among cloud textures, using only raw multispectral imagery as input. Cloud classes are therefore defined based on spectral properties and spatial textures without reliance on location, time/season, derived physical properties, or pre-designated class definitions. We use this approach to generate a unique new cloud dataset, the AI-driven cloud classification atlas (AICCA), which clusters 22 years of ocean images from the Moderate Resolution Imaging Spectroradiometer (MODIS) on NASA’s Aqua and Terra instruments—198 million patches, each roughly 100 km × 100 km (128 × 128 pixels)—into 42 AI-generated cloud classes, a number determined via a newly-developed stability protocol that we use to maximize richness of information while ensuring stable groupings of patches. AICCA thereby translates 801 TB of satellite images into 54.2 GB of class labels and cloud top and optical properties, a reduction by a factor of 15,000. The 42 AICCA classes produce meaningful spatio-temporal and physical distinctions and capture a greater variety of cloud types than do the nine International Satellite Cloud Climatology Project (ISCCP) categories—for example, multiple textures in the stratocumulus decks along the West coasts of North and South America. We conclude that our methodology has explanatory power, capturing regionally unique cloud classes and providing rich but tractable information for global analysis. AICCA delivers the information from multi-spectral images in a compact form, enables data-driven diagnosis of patterns of cloud organization, provides insight into cloud evolution on timescales of hours to decades, and helps democratize climate research by facilitating access to core data.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Multi-Source Machine Learning and Thermoplastics Enhanced Aerostructure Manufacturing (mTEAM)

RTX Technology Research Center (RTRC), together with Collins Aerospace (Collins) and Oak Ridge National Laboratory (ORNL) has developed an Artificial Intelligence (AI) / Machine Learning (ML) guided solution to advance the manufacturing and assembly of high performance and lightweight thermoplastic composite (TPC) aerospace products. The solution aims to lower risk, cost and lead time for induction heating based welding and consolidation processes for TPC structure. The cost and lead time of part and material specific process development for induction welding (IW) and induction consolidation will be reduced by replacing traditional empirical methods with optimization methods that merge AI/ML and physics-based process simulations and process experiments with sensing and controls. TPC-IW process development is empirical in nature, and uncertainties in material & process behavior exist near & far from the induction coil. Physics-based simulations can be leveraged directly for process optimization but can be too computationally expensive to run in high fidelity and real time to do robust process optimization. The key impact of successful TPC induction consolidation and welding is cost & lead time reduction for part & material specific consolidation and welding recipes. This is an enabler for more rapid deployment of TPC structures via joining assembly, which can reduce energy & cost intensive usage of autoclaves & ovens. The solution aimed to advance the U.S. Department of Energy’s interests in using thermoplastics and automation in composite manufacturing for improvement of products for existing markets via increased production speeds, reduced costs, and lowered use of energy. Welded TPC structures can offer significant weight & energy savings for high-value commercial aerospace & industrial applications compared to metal & thermoset composite structures assembled by mechanical fastening and/or adhesive bonding. The project was organized into two Budget Periods. Budget Period 1 (BP1) was 15 months and its goal was to perform ML process optimization framework development & deployment on lab-coupon aerostructure components. A Go/No-Go Review was performed at the end of BP1 to verify fulfilment of key tasks & milestones to justify a Go Decision to move into the next Budget Period. Budget Period 2 (BP2) was 12 months and its goal was the deployment of the ML framework for ML process optimization of pilot industrial scale aerostructure components. The overall project aim was to develop & demonstrate ML-enhanced modeling framework that learns process-property mapping from multiple data sources at different fidelities. During BP1, the team accomplished key tasks & milestones to demonstrate the concept of multi-source ML for TPC aerostructure consolidation and assembly. First, the team completed documentation of induction based TPC heating requirements including baseline metrics to compare measured results against. Next the team completed demonstration of data generation from physics-based simulations for ML surrogate model generation and demonstrated the integration of physics-based simulation data into multi-source AI/ML algorithms. In parallel, the team established the lab-coupon scale induction welding system and completed a process to label and reduce generated data from physics-based simulation and experiments for ML surrogate models to enable multi-source ML model training & testing. To complete BP1, the team integrated physics-based simulation data and experimental data into multi-source ML algorithms. This was based on the team completing ML deployment of the induction welding on a lab system at RTRC and AI/ML deployment on existing induction welding line at Collins. ORNL visited both Collins and RTRC sites to witness the TPC induction welding process. Then, ORNL designed and constructed a new version of their vision-based sensing system better adapted to acquire process signals of the TPC induction welding process for process anomaly and defect detection. In BP2, the team accomplished key tasks & milestones to scale up multi-source ML for TPC aerostructure consolidation and assembly from the lab-coupon scale to the pilot-industrial scale. In BP2, the team demonstrated real time anomaly & defect detection via experiments performed by ORNL & RTRC. The team completed ML-optimization heating trials for TPC induction consolidation at Collins, and the team confirmed pilot industrial scale experimental data from Collins was compatible with the developed ML pipeline from RTRC. The team completed sub-element scale ML process optimization demonstration at RTRC, where the team leveraged RTRC’s robotic TPC welding setup to de-risk the ML process optimization by performing ML analysis of recorded temperatures to account for complex part features. Then, the team applied its ML-derived control strategies and ML process optimization framework at Collins to the pilot-industrial scale on a demo skin-stiffener part representative of a nacelle aerostructure fan cowl section. The key innovation is the AI/ML framework enabling effective process development of high performance, lightweight, energy efficient TPCs for composite aircraft structures.

36 MATERIALS SCIENCE↗
DOE OSTI2023

MnEdgeNet for accurate decomposition of mixed oxidation states for Mn XAS and EELS L2,3 edges without reference and calibration

Accurate decomposition of the mixed Mn oxidation states is highly important for characterizing the electronic structures, charge transfer and redox centers for electronic, and electrocatalytic and energy storage materials that contain Mn. Electron energy loss spectroscopy (EELS) and soft X-ray absorption spectroscopy (XAS) measurements of the Mn L2,3 edges are widely used for this purpose. To date, although the measurements of the Mn L2,3 edges are straightforward given the sample is prepared properly, an accurate decomposition of the mix valence states of Mn remains non-trivial. For both EELS and XAS, 2+, 3+, and 4+ reference spectra need to be taken on the same instrument/beamline and preferably in the same experimental session because the instrumental resolution and the energy axis offset could vary from one session to another. To circumvent this hurdle, in this study, we adopted a deep learning approach and developed a calibration-free and reference-free method to decompose the oxidation state of Mn L2,3 edges for both EELS and XAS. A deep learning regression model is trained to accurately predict the composition of the mix valence state of Mn. To synthesize physics-informed and ground-truth labeled training datasets, we created a forward model that takes into account plural scattering, instrumentation broadening, noise, and energy axis offset. With that, we created a 1.2 million-spectrum database with 1-by-3 oxidation state composition ground truth vectors. The library includes a sufficient variety of data including both EELS and XAS spectra. By training on this large database, our convolutional neural network achieves 85% accuracy on the validation dataset. We tested the model and found it is robust against noise (down to PSNR of 10) and plural scattering (up to t/λ = 1). We further validated the model against spectral data that were not used in training. In particular, the model shows high accuracy and high sensitivity for the decomposition of Mn 3 O 4 , MnO, Mn 2 O 3 , and MnO 2 . The accurate decomposition of Mn 3 O 4 experimental data shows the model is quantitatively correct and can be deployed for real experimental data. Our model will not only be a valuable tool to researchers and material scientists but also can assist experienced electron microscopists and synchrotron scientists in the automated analysis of Mn L edge data.

25 ENERGY STORAGE↗
DOE OSTI2021

DETECTING FIRE WITH MACHINE LEARNING-ENABLED VISUAL MONITORING FOR NUCLEAR POWER PLANT ENVIRONMENTS

Nuclear power plants are experiencing significant cost challenges to remain competitive with other energy-generation utilities. Unlike other industries, the cost of operation and maintenance activities is mostly attributed to workforce costs. To mitigate this, nuclear power plant stakeholders are increasingly interested in the development and deployment of machine learning methods to potentially automate or augment manually intensive tasks to reduce costs, especially for monitoring activities. One monitoring function that is visually demanding and that can occur frequently to meet the requirements of a fire protection program is visually monitoring an area for fire occurrence. Currently, fire watch activities consist of a worker physically stationed at a given location with the sole responsibility of observing a given area to ensure a fire is detected and mitigated promptly. This effort focused on the development and evaluation of a suitable deep convolutional neural network to classify individual video frames at a sub-second frequency for the occurrence of “fire” and “no fire” in varying industrial environments similar to nuclear power plants. It is believed that a trained neural network model could be integrated with existing facility video surveillance camera feeds to generate alerts when fire inferences occur in individual frames captured at sub-second temporal resolutions. Extensive effort was dedicated to identifying and curating suitable imagery training data representing varying environments and scene settings with and without flame features to maximize generalization in nuclear power plant environments. The data collection effort resulted in the aggregation of a large, labeled image library exceeding 12,000 images to support model training for diverse industrial environments. A deep neural network model incorporating parallel multi-scale capabilities was developed and trained to support accurate image-based detection of flame incidents of varying sizes and spectral feature properties within heterogeneous scenes. Analysis results show that the trained model can achieve high inference accuracy despite heterogeneous scene environments and components. Testing accuracy exceeded 95.0 percent with very low false positive and false negative inferences.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2024

Urban morphology and urban water demand evolution in the Los Angeles region

Detailed description of the dataset sources used in this study, the experimental workflow, and plotting for the paper figures provided at the associated GitHub Meta Repo: https://github.com/IMMM-SFA/Ferencz_et_al_2024_ERL The future water demand projections from this study are hypothetical future water demands that reflect the population and urban land cover changes represented by the scenarios considered. The intent and emphasis of this work is investigating the interactions between population change, evolution of urban morphology, and water demand. These projections are not meant to be likely future demands for specific water providers or the LA region and should not be interpreted as such. The folders contain input and output data for each step of the "Recreate my Experiment" workflow described in the associated GitHub meta-repository as well as data used for plotting Figures for the paper that this dataset supports. Description of each folder's contents and use: Step_1a: All necessary inputs to the associated python script provided on the GitHub repo. Step_1b: All necessary inputs (downscaled population rasters) used by the associated python script provided on the GitHub repo. Original 1-km squared rasters that were downscaled also provided. Step_1c: Urban growth projection rasters corresponding to SSP3 and SSP5 population scenarios are provided in separate subfolders as well as the water provider boundaries used for analysis. Outputs of data processing also provided. Associated python script provided on GitHub. Step_1d: Description of Inputs used by the QGIS Model Builder GUI that automates geospatial processing and clipping the of the high resolution land cover data for each urban land class footprint within a defined polygon boundary. The Model Builder is provided on the GitHub repo and can be used by QGIS. The outputs of this step are in "Clipped Provider Hi Res Landcover". If the user wants to use The Model Builder for different regions of LA or two test our outputs, they will need to download the hi resolution landcover raster listed in the Readme and in Ref [2] of the GitHub Page. Step_1e: All necessary inputs to generate average monthly demand for each water provider. Associated python script on GitHub. Step 2: Output data about land cover metrics (areas and fractions) for each urban land class for each water provider. Associated python script on GitHub. Uses outputs from Step 1d "Clipped Provider Hi Res Landcover" Step 3: Inputs for and Outputs from the urban projection raster analysis Python script on GitHub. The outputs are rasters of urban pixels that were converted to a higher land class and the number of land class units that changed (Values of 1, 2, or 3). For example, a value of 2 could be LC 21 -> 23 or LC 22 -> 24. These maps are label "intensification." The other outputs are "urban growth" rasters showing the conversion of non urban to urban land, which are indicated by pixel values of 1. Step 4: Output projections of indoor and outdoor annual and monthly demands for each water provider. These are used for Figures 4 - 7 of the paper. Figures: This folder has data used for plotting Figures 1 through 5. Data for Figures 6 and 7 are sourced directly from folders associated with the Processing and Analysis Steps 1 - 4 and the plotting scripts for Figures 6 and 7 are commented with what folder paths are needed to generate the figures. The GitHub page provides descriptions of how each figure was made and the associated plotting scripts used.

Los Angeles↗
DOE OSTI2023

Artificial intelligence models, photos, and data associated with the manuscript “Quantifying Streambed Grain Size, Uncertainty, and Hydrobiogeochemical Parameters Using Machine Learning Model YOLO” (v2)

This data package is associated with the manuscript “Quantifying Streambed Grain Size, Uncertainty, and Hydrobiogeochemical Parameters Using Machine Learning Model YOLO” published in Water Resources Research (Chen et al., 2024). This data package includes the training, validation, testing, and prediction data used by the artificial intelligence (AI) model for automated grain size and hydro-biogeochemistry quantification using streambed photos. The grain size data are extracted for each photo using You Look Only Once (YOLO), a pre-trained object detection model. This data package was originally published in October 2023. It was updated August 2025 (v2; new and modified files). File and folder names were not revised to indicate changes. See the change history section in the readme for more details. Please see flmd.csv for a list of all files contained in this data package and descriptions for each. Please see dd.csv for a data dictionary that defines the column headers of .csv files in the data package. This dataset is comprised of one data folder containing (1) file-level metadata; (2) data dictionary; (3) readme; and (4) six subfolders. Subfolders 1 to 4 include the training, validation, testing, and prediction data. Subfolder 5_Summary includes the summary results of different combinations of training, validation, testing, and prediction data. Subfolder 6_SupplementalData includes additional data downloaded from public sources (Kaufman et al., 2023a; Kaufman et al., 2023b; Garefalakis et al., 2023; Mair et al., 2024; https://github.com/river-corridors-sfa/Geospatial_variables). In total, the data package includes 110 folders and 44,283 files. These files include 9,047 .jpg photos, 1 .png photo, 3 .tif photos; 26,639 photo labels and individual grain sizes and probability from AI (.txt); 8,447 grain size distribution data (.dat); and 126 CSV files for results summary, and 14 required metadata files (.xlsx). The summary CSV files contain 68 columns and approximately 2,200 rows that represent photo names, site locations, recording time, GPS coordinates, grains sizes (D10, D50, D60, and D84), number of grains, and additional hydro-biogeochemical data such as water depth, flow velocity, Manning’s coefficient, friction factor, hydraulic conductivity, permeability, streambed interstitial velocity magnitude, mass transfer rate, and nitrate uptake velocity. The photos were obtained from 75 sites in the Yakima River Basin and the Columbia River shorelines, and other associated data from samples and sensors obtained when the photos were taken are publicly available (Fulton et al. 2022; Grieger et al. 2023). All files are .csv, .txt, .dat, .jpg, or .pdf. We acknowledge the Yakama Nation as owners and caretakers of the lands where we collected some of these data. We thank the Confederated Tribes and Bands of the Yakama Nation Tribal Council and Yakama Nation Fisheries for working with us to facilitate sample collection and optimization of data usage according to their values and worldview.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

Deep learning uncertainty quantification for clinical text classification

Machine learning algorithms are expected to work side-by-side with humans in decision-making pipelines. Thus, the ability of classifiers to make reliable decisions is of paramount importance. Deep neural networks (DNNs) represent the state-of-the-art models to address real-world classification. Although the strength of activation in DNNs is often correlated with the network’s confidence, in-depth analyses are needed to establish whether they are well calibrated. In this paper, we demonstrate the use of DNN-based classification tools to benefit cancer registries by automating information extraction of disease at diagnosis and at surgery from electronic text pathology reports from the US National Cancer Institute (NCI) Surveillance, Epidemiology, and End Results (SEER) population-based cancer registries. In particular, we introduce multiple methods for selective classification to achieve a target level of accuracy on multiple classification tasks while minimizing the rejection amount—that is, the number of electronic pathology reports for which the model’s predictions are unreliable. We evaluate the proposed methods by comparing our approach with the current in-house deep learning-based abstaining classifier. Overall, all the proposed selective classification methods effectively allow for achieving the targeted level of accuracy or higher in a trade-off analysis aimed to minimize the rejection rate. On in-distribution validation and holdout test data, with all the proposed methods, we achieve on all tasks the required target level of accuracy with a lower rejection rate than the deep abstaining classifier (DAC). Interpreting the results for the out-of-distribution test data is more complex; nevertheless, in this case as well, the rejection rate from the best among the proposed methods achieving 97% accuracy or higher is lower than the rejection rate based on the DAC. We show that although both approaches can flag those samples that should be manually reviewed and labeled by human annotators, the newly proposed methods retain a larger fraction and do so without retraining—thus offering a reduced computational cost compared with the in-house deep learning-based abstaining classifier.

59 BASIC BIOLOGICAL SCIENCES↗