Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “label quality”

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 OSTI2022

Stable parallel training of Wasserstein conditional generative adversarial neural networks

In this work, we propose a stable, parallel approach to train Wasserstein conditional generative adversarial neural networks (W-CGANs) under the constraint of a fixed computational budget. Differently from previous distributed GANs training techniques, our approach avoids inter-process communications, reduces the risk of mode collapse and enhances scalability by using multiple generators, each one of them concurrently trained on a single data label. The use of the Wasserstein metric also reduces the risk of cycling by stabilizing the training of each generator. We illustrate the approach on the CIFAR10, CIFAR100, and ImageNet1k datasets, three standard benchmark image datasets, maintaining the original resolution of the images for each dataset. Performance is assessed in terms of scalability and final accuracy within a limited fixed computational time and computational resources. To measure accuracy, we use the inception score, the Fréchet inception distance, and image quality. An improvement in inception score and Fréchet inception distance is shown in comparison to previous results obtained by performing the parallel approach on deep convolutional conditional generative adversarial neural networks as well as an improvement of image quality of the new images created by the GANs approach. Weak scaling is attained on both datasets using up to 2000 NVIDIA V100 GPUs on the OLCF supercomputer Summit.

97 MATHEMATICS AND COMPUTING↗
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 OSTI2022

Mass Detection for Heavy-Duty Vehicles using Gaussian Belief Propagation

Predicting vehicle mass is critical to accurately estimate energy use and emissions of commercial trucks. However, data from vehicle telematics is often not at sufficient temporal resolution or accuracy for use in model-based detection methods. In this work, a new statistical mass prediction technique is described for heavy-duty vehicles that incorporates the use Gaussian Belief Propagation (GBP) for probabilistic inference. Similar to Bayesian inference models, the GBP model typically requires less labeled training data than other contemporary machine learning techniques. First, a factor graph is constructed, and a set of Gaussian belief nodes with associated means and variances are fitted to the training data. To better handle noisy input data, the GBP mass prediction model utilizes a k-nearest factors (kNF) algorithm for probabilistic inference on unseen testing data. The proposed method is compared with a classical weighted k-nearest neighbors (kNN) regressor. This statistical kNF-GBP model works even with low-quantity, low-quality initial training data, while being capable of realtime mass estimation. Unlike the kNN regressor, the GBP model produces a measure of uncertainty with its predictions. The proposed method is validated using curve-sampled driving data collected from multiple cloud-connected Class 8 regional haul diesel trucks. Both the kNN regressor and the kNF-GBP mass prediction model were able to predict payload mass with coefficients of determination above 0.97 with minimal data preprocessing.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗
DOE OSTI2023

DeepAstroUDA: semi-supervised universal domain adaptation for cross-survey galaxy morphology classification and anomaly detection

Abstract Artificial intelligence methods show great promise in increasing the quality and speed of work with large astronomical datasets, but the high complexity of these methods leads to the extraction of dataset-specific, non-robust features. Therefore, such methods do not generalize well across multiple datasets. We present a universal domain adaptation method, DeepAstroUDA , as an approach to overcome this challenge. This algorithm performs semi-supervised domain adaptation (DA) and can be applied to datasets with different data distributions and class overlaps. Non-overlapping classes can be present in any of the two datasets (the labeled source domain, or the unlabeled target domain), and the method can even be used in the presence of unknown classes. We apply our method to three examples of galaxy morphology classification tasks of different complexities (three-class and ten-class problems), with anomaly detection: (1) datasets created after different numbers of observing years from a single survey (Legacy Survey of Space and Time mock data of one and ten years of observations); (2) data from different surveys (Sloan Digital Sky Survey (SDSS) and DECaLS); and (3) data from observing fields with different depths within one survey (wide field and Stripe 82 deep field of SDSS). For the first time, we demonstrate the successful use of DA between very discrepant observational datasets. DeepAstroUDA is capable of bridging the gap between two astronomical surveys, increasing classification accuracy in both domains (up to 40 % on the unlabeled data), and making model performance consistent across datasets. Furthermore, our method also performs well as an anomaly detection algorithm and successfully clusters unknown class samples even in the unlabeled target dataset.

79 ASTRONOMY AND ASTROPHYSICS↗
NASA NTRS1983

User's guide for the Solar Backscattered Ultraviolet (SBUV) and the Total Ozone Mapping Spectrometer (TOMS) RUT-S and RUT-T data sets: October 31, 1978 to November 1, 1980

Raw data from the Solar Backscattered Ultrviolet/Total Ozone Mapping Spectrometer (SBUV/TOMS) Nimbus 7 operation are available on computer tape. These data are contained on two separate sets of RUTs (Raw Units Tapes) for SBUV and TOMS, labelled RUT-S and RUT-T respectively. The RUT-S and RUT-T tapes contain uncalibrated radiance and irradiance data, housekeeping data, wavelength and electronic calibration data, instrument field-of-view location and solar ephemeris information. These tapes also contain colocated cloud, terrain pressure and snow/ice thickness data, each derived from an independent source. The "RUT User's Guide" describes the SBUV and TOMS experiments, the instrument calibration and performance, operating schedules, and data coverage, and provides an assessment of RUT-S and -T data quality. It also provides detailed information on the data available on the computer tapes.

Fleig, A. J.↗
DOE OSTI2022

Machine learning-assisted upscaling analysis of reservoir rock core properties based on micro-computed tomography imagery

Optimum solutions for geologic modeling and reservoir simulation in industries such as oil and gas recovery and carbon capture and storage require accurate characterization of reservoir properties, which are often heterogeneous. In this study, high-quality micro-computed tomography (CT) images (1.475-μm/pixel resolution) of a sandstone core acquired from the Bell Creek oil field, USA, were used to provide nondestructive analysis of pore- and core-scale heterogeneity across measurement scales of 94–566 μm. In addition to characterizing the as-received sample, the core sample was flooded with brine to evaluate the capacity of the core sample to receive injected fluids. The micro-CT images were systematically segmented into pore spaces and grains via machine learning (ML) steps including image preprocessing, label creation using a traditional ML method based on limited manual image annotation, and finally U-Net segmentation. The segmented image stacks were reconstructed into digital cubes of various scales of voxel lengths. The 3D porosity values were calculated for all the digital cubes, and the fractal dimensions of the cubes were estimated using a box-counting method. The results showed that smaller cubes had greater heterogeneity and that the porosity values could be accurately estimated by fractal dimension and voxel lengths using ML models. For the core sample with brine flooding, the ratio of pores filled by brine to the total pore space was related to the porosity and could also be accurately estimated by porosity, fractal dimension, and voxel lengths using ML models. In conclusion, the results of this study demonstrate that the concept of fractal dimension can be a useful vector to perform upscaling analysis of sandstone rock heterogeneity from the pore to core scale and that fractal dimensions can be used to estimate porosity values and pore space-filling capacity across those scales.

58 GEOSCIENCES↗
NASA NTRS2018

Enriching the Twitter Stream Increasing Data Mining Yield and Quality Using Machine Learning

Social media data streams are important sources of real-time and historical global information for science applications. At the NASA Goddard Earth Sciences Data and Information Services Center (GES DISC), we are exploring the Twitter data stream for its potential in augmenting the validation program of NASA Earth science missions, specifically the Global Precipitation Measurement (GPM) mission. We have implemented a tweet processing infrastructure that outputs classified precipitation tweets. Inputs are "passive" tweets, along with a smaller number of tweets from "active" participants, i.e., those knowingly contributing to our effort. The "active" tweets, presumably of higher quality, enrich the Twitter stream. "Active" sources include data scraped from other social media (e.g., public Facebook posts) and data from existing crowdsourcing programs (e.g., mPING reports). In addition, there is likely relevant precipitation information in images and documents that are the end points of links often included in tweets. Information derived from these "active" sources could then be tweeted into the Twitter stream, thus enriching its quality. The objective of our current work is to mine these tweet­ linked images and documents, using neural networks, to increase the information content and quality related to precipitation. For images, we classified them as either precipitation-related or not. For training and validation, we used images obtained via the Google custom search API. We created two models: (1) by training a simple Convolutional Neural Network and (2) by using transfer learning principles to adapt a pre-trained object recognition model. For documents, both those linked to tweets and the tweet contents, we trained Hierarchical Attention Networks to determine precipitation occurrence, type, and intensity. For training and validation, we used a keyword-filtered tweet data set labelled with ground truth data from Dark Sky (an API to retrieve weather-related labels) and the National Severe Storms Laboratory's Multi­ Radar/Multi-Sensor (MRMS) system. Our results demonstrated the efficacy of our machine learning approaches for enriching the Twitter stream, to derive information potentially useful for validation of earth science satellite data.

Albayrak, Arif↗
NASA NTRS2012

Evaluation of Enhanced Low Dose Rate Sensitivity in Discrete Bipolar Junction Transistors

We evaluate the low dose rate sensitivity in several families of discrete bipolar transistors across device parameter, quality assurance level, and irradiation bias configuration. The 2N2222 showed the most significant low dose rate sensitivity, with low dose rate enhancement factor of 3.91 after 100 krad(Si). The 2N2907 also showed critical degradation levels. The devices irradiated at 10 mrad(Si)/s exceeded specifications after 40 and 50 krad(Si) for the 2N2222 and 2N2907 devices, respectively.

Chen, Dakai↗
DOE OSTI2021

PSpecteR: A User-Friendly and Interactive Application for Visualizing Top-Down and Bottom-Up Proteomics Data in R

Visual examination of mass spectrometry data is necessary to assess data quality and to facilitate data exploration. Graphics provide the means to evaluate spectral properties, test alternative peptide/protein sequence matches, prepare annotated spectra for publication, and fine-tune parameters during wet lab procedures. Visual inspection of MS data is hindered by proprietary proteomics visualization software designed for particular workflows and academic software that lack visualization tools. We built PSpecteR, an open-source and interactive R Shiny web application to address these issues, with support for several steps of proteomics data processing, including: reading various mass spectrometry files, running open-source database search tools, labelling spectra with fragmentation patterns, testing post-translational modifications, plotting where identified fragments map to reference sequences, and visualizing algorithmic output and metadata. All figures, tables, and spectra are exportable within one easy-to-use graphical user interface. Our current software provides a flexible and modern R framework to support fast implementation of additional features. The open source code is readily available (https://github.com/EMSL-Computing/PSpecteR), and a PSpecteR Docker container (https://hub.docker.com/r/emslcomputing) is available for easy local installation.

96 KNOWLEDGE MANAGEMENT AND PRESERVATION↗
DOE OSTI2022

Multiview Incomplete Knowledge Graph Integration with application to cross-institutional EHR data harmonization

Objective: The growing availability of electronic health records (EHR) data opens opportunities for integrative analysis of multi-institutional EHR to produce generalizable knowledge. A key barrier to such integrative analyses is the lack of semantic interoperability across different institutions due to coding differences. We propose a Multiview Incomplete Knowledge Graph Integration (MIKGI) algorithm to integrate information from multiple sources with partially overlapping EHR concept codes to enable translations between healthcare systems. Methods: The MIKGI algorithm combines knowledge graph information from (i) embeddings trained from the co-occurrence patterns of medical codes within each EHR system and (ii) semantic embeddings of the textual strings of all medical codes obtained from the Self-Aligning Pretrained BERT (SAPBERT) algorithm. Due to the heterogeneity in the coding across healthcare systems, each EHR source provides partial coverage of the available codes. MIKGI synthesizes the incomplete knowledge graphs derived from these multi-source embeddings by minimizing a spherical loss function that combines the pairwise directional similarities of embeddings computed from all available sources. MIKGI outputs harmonized semantic embedding vectors for all EHR codes, which improves the quality of the embeddings and enables direct assessment of both similarity and relatedness between any pair of codes from multiple healthcare systems. Results: With EHR co-occurrence data from Veteran Affairs (VA) healthcare and Mass General Brigham (MGB), MIKGI algorithm produces high quality embeddings for a variety of downstream tasks including detecting known similar or related entity pairs and mapping VA local codes to the relevant EHR codes used at MGB. Based on the cosine similarity of the MIKGI trained embeddings, the AUC was 0.918 for detecting similar entity pairs and 0.809 for detecting related pairs. For cross-institutional medical code mapping, the top 1 and top 5 accuracy were 91.0% and 97.5% when mapping medication codes at VA to RxNorm medication codes at MGB; 59.1% and 75.8% when mapping VA local laboratory codes to LOINC hierarchy. When trained with 500 labels, the lab code mapping attained top 1 and 5 accuracy at 77.7% and 87.9%. MIKGI also attained best performance in selecting VA local lab codes for desired laboratory tests and COVID-19 related features for COVID EHR studies. Compared to existing methods, MIKGI attained the most robust performance with accuracy the highest or near the highest across all tasks. Conclusions: The proposed MIKGI algorithm can effectively integrate incomplete summary data from biomedical text and EHR data to generate harmonized embeddings for EHR codes for knowledge graph modeling and cross-institutional translation of EHR codes.

Zhou, Doudou↗
DOE OSTI2020

Transport of intensity equation: a tutorial

When it comes to “phase measurement” or “quantitative phase imaging”, many people will automatically connect them with “laser” and “interferometry”. Indeed, conventional quantitative phase imaging and phase measurement techniques generally rely on the superposition of two beams with a high degree of coherence: complex interferometric configurations, stringent requirements on the environmental stabilities, and associated laser speckle noise severely limit their applications in optical imaging and microscopy. On a different note, as one of the most well-known phase retrieval approaches, the transport of intensity equation (TIE) provides a new non-interferometric way to access quantitative phase information through intensity only measurement. Despite the insufficiency for interferometry, TIE is applicable under partially coherent illuminations (like the Köhler’s illumination in a conventional microscope), permitting optimum spatial resolution, higher signal-to-noise ratio, and better image quality. In this tutorial, we give an overview of the basic principle, research fields, and representative applications of TIE, focus particularly on optical imaging, metrology, and microscopy. The purpose of this tutorial is twofold. It should serve as a self-contained introduction to TIE for readers with little or no knowledge of TIE. On the other hand, it attempts to give an overview of recent developments in this field. These results highlight a new era in which strict coherence and interferometry are no longer prerequisites for quantitative phase imaging and diffraction tomography, paving the way toward new generation label-free three-dimensional microscopy, with applications in all branches of biomedicine.

42 ENGINEERING↗
DOE OSTI2021

Automated theoretical chemical kinetics: Predicting the kinetics for the initial stages of pyrolysis

Large scale implementation of high level computational theoretical chemical kinetics offers the prospect for dramatically improving the fidelity of combustion chemical modeling. To facilitate such efforts, we have developed a suite of codes, collectively referred to as AutoMech, that allow for the automatic prediction of the kinetics for large sets of reactions via ab initio transition-state-theory based master-equation calculations. The primary input is simply the mechanism, a dictionary relating chemically identifiable species descriptors (e.g., SMILES or InChIs) to species labels in the mechanism, and a specification of the electronic structure and transition state theory models to be implemented. Here we illustrate the current utility of AutoMech through a study of the initial stages of pyrolysis for 3 sets of fuels: sequences of alkanes, alcohols, and aldehydes. For simplicity, the analysis focuses on abstractions from the fuel by H, CH 3 , and OH, and the decomposition of the resulting radicals. Altogether, there are a total of 166 input channels in these sets (more than 363 forward reactions when expanded to the full set of elementary reactions). The code successfully produces high quality rate estimates (with apparent uncertainties less than a factor of two in limited comparisons with experiment) for > 95% of these. For the radical decomposition reactions, the analysis includes predictions for the pressure dependence of the kinetics. This wide-ranging exploration illustrates (i) the effect of different levels of prediction on the expected accuracy, (ii) the branching between abstractions at different sites for different abstractors, (iii) the dependence of the rates on the chemical structure, and (iv) the variation in radical stabilities across chemical families. These results, as well as the demonstrated feasibility of the methodology, should find further utility in the development of accurate rate expressions for arbitrary fuels. (c) 2020 The Combustion Institute. Published by Elsevier Inc. All rights reserved.

Elliott, Sarah N.↗
DOE OSTI2021

Probing Light-Dependent Regulation of the Calvin Cycle Using a Multi-Omics Approach

Photoautotrophic microorganisms are increasingly explored for the conversion of atmospheric carbon dioxide into biomass and valuable products. The Calvin-Benson-Bassham (CBB) cycle is the primary metabolic pathway for net CO 2 fixation within oxygenic photosynthetic organisms. The cyanobacteria, Synechocystissp. PCC 6803, is a model organism for the study of photosynthesis and a platform for many metabolic engineering efforts. The CBB cycle is regulated by complex mechanisms including enzymatic abundance, intracellular metabolite concentrations, energetic cofactors and post-translational enzymatic modifications that depend on the external conditions such as the intensity and quality of light. However, the extent to which each of these mechanisms play a role under different light intensities remains unclear. In this work, we conducted non-targeted proteomics in tandem with isotopically non-stationary metabolic flux analysis (INST-MFA) at four different light intensities to determine the extent to which fluxes within the CBB cycle are controlled by enzymatic abundance. The correlation between specific enzyme abundances and their corresponding reaction fluxes is examined, revealing several enzymes with uncorrelated enzyme abundance and their corresponding flux, suggesting flux regulation by mechanisms other than enzyme abundance. Additionally, the kinetics of 13 C labeling of CBB cycle intermediates and estimated inactive pool sizes varied significantly as a function of light intensity suggesting the presence of metabolite channeling, an additional method of flux regulation. These results highlight the importance of the diverse methods of regulation of CBB enzyme activity as a function of light intensity, and highlights the importance of considering these effects in future kinetic models.

54 ENVIRONMENTAL SCIENCES↗
NASA NTRS1993

Gas-rich dwarf galaxies in dense and sparse environments

Dwarf irregular galaxies (generically labelled Im for the present purposes) pose an enigma to students of galaxy evolution. In nearby groups and the Virgo cluster, Im galaxies are at least as abundant as spiral galaxies, and their low surface brightnesses and high gas-to-stars ratios suggest that (at least in the stochastic self-propagating star formation scenario) there should be significant numbers of HI clouds with masses approaching 10(exp 8) solar mass which have undergone very little or no star formation. To date, however, no clouds with so little star formation that they would not be recognized as Im galaxies on high-quality photographic plates have been identified. There have been suggestions that such dwarfs may be tidally disrupted in regions of high galactic density, but may be prevalent in low density regions. We offer data from three parallel programs relevant to this issue. (1) A large number of Im galaxies throughout the Local Supercluster have been mapped in the HI spectral line using the Arecibo Radiotelescope, and we can establish the frequency with which HI disks much more extended than their optically visible portions are found. (2) Our extensive mapping of spiral and dwarf galaxies in the Virgo cluster allows us to set stringent limits on the density of star-free Hi clouds in that cluster. (3) We have conducted a sampling of the void in the distribution of galaxies toward the super galactic pole, optimized for finding low-mass HI clouds at redshifts out to approximately 2000 km/s.

Hoffman, G. Lyle↗
DOE OSTI2025

Lignin Removal in Subcellular Location of Poplar Cell Wall During Pretreatment Significantly Impacts Cellulose Digestibility

The γ-valerolactone (GVL) pretreatment is one of the leading solvent-based methods for producing high-quality lignin under mild conditions. However, the glucan conversion yield from GVL pretreated biomass remains unsatisfactory. To explore the discrepancies between the relatively low glucan conversion and high lignin extraction, we conducted GVL−HCl and NaOH pretreatments on poplar and investigated their effects on lignin content and location, as well as on enzymatic hydrolysis of poplar cell walls at the subcellular level. Under designated pretreatment conditions of GVL−HCl (90% GVL, 0.1 M HCl, 100 °C, 1 h) and NaOH (1 M, 121 °C, 2 h), the glucan conversion yields were 69.4% and 95.8%, with lignin removal rates of 67.8% and 47.7%, respectively. Four types of GFP-labeled carbohydrate binding modules were used to identify different forms of cellulose in the pretreated cell walls. The overall binding intensities to pretreated poplar were stronger for NaOH compared to GVL−HCl pretreatment. Stimulated Raman scattering microscopy imaging revealed that GVL−HCl preferentially extracted lignin from the compound middle lamella and cell corner areas, while NaOH effectively dissolved lignin in the secondary cell walls. Real-time imaging of cellulase degradation of pretreated cell walls further indicated that digestion started from both the cell lumen and the compound middle lamella areas for GVL, whereas it occurred uniformly across the secondary cell walls for NaOH. Our findings suggest that the location of lignin removal during pretreatment is crucial for enzymatic cellulose degradation, in addition to the total amount of lignin extraction.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2021

Phasor-Measurement-Unit-Based Data Analytics Using Digital Twin and PhasorAnalytics Software

A major objective of this project was to apply GE’s commercial machine learning and data analytics toolsets to large-scale, real-world, anonymized Phasor Measurement Unit (PMU) datasets in order to extract signatures, correlated and/or causal factors, and precursor patterns associated with significant power system phenomena. The project had a particular emphasis on extraction of insights relevant to asset health monitoring, real-time load modeling and cybersecurity monitoring. Additionally, the team was directed to undertake a comprehensive data quality analysis for the provided datasets and encouraged to estimate the ‘machine-learning readiness’ of the datasets by documenting any major obstacles to the application of commercial machine learning algorithms. To accomplish the aforementioned objectives, the project team’s work centered around the identification of key event signatures and application of the identified event signatures for event detection and event classification. The industry-validated, semi-supervised machine learning strategy employed for event signature identification involved several major tasks, including data-preprocessing, generation of an overabundance of features, normal data identification, normality modeling, and event signature identification through a methodical, quantitative ranking of features in order of relevance to each studied event type. Throughout the project, data quality issues and mitigation techniques were investigated. In this report, insights are provided regarding the readiness of the provided synchrophasor datasets for application of machine learning and data analytics. The methodologies employed for this technical strategy are summarized in this report. With regards to data preprocessing and feature generation, the provided Training and Test Datasets were ingested into GE’s big data environment. Subsequently, the team applied bad data cleansing and data imputation scripts, event detection scripts, and application programming interfaces (APIs) to the datasets for convenient data access. The project team completed development and validation of dozens of physics-based, statistics-based and transformation-based feature functions used for the extraction of over 60 synchrophasor features. Using a new parallel feature generation technology developed on this project, over 60 features have been rapidly generated for the full two years’ worth of Training and Test Dataset data associated with both the Eastern and Western interconnects. Even accommodating for temporal down-sampling inherent to the feature extraction procedure, this parallel feature generation activity resulted in a massive feature set with a storage requirement approximately equal to that of the raw training dataset itself. With regards to normal data identification and normality modeling, a normality model was built using the feature data extracted from the Training Dataset and iteratively refined subsequent to incremental adjustments and expansions of the Training Dataset feature data. With respect to event characterization and signature identification, an event signature identification pipeline was developed and used in conjunction with the normality model to identify over 15 event signatures for key event categories within the Training Dataset. The identified event signatures were used to characterize hundreds of key events in terms of relative severity, duration, and location of the event. An investigation was undertaken to identify correlated and causal factors involved in transformer events. A separate investigation into temporal trends in ring-down analysis results was undertaken to determine possible associations between system dynamics and various other factors such as loading, season or year. To validate the identified event signatures, additional work was undertaken to develop signature-based anomaly detection and classification tools suitable for convenient application to the synchrophasor datasets. The anomaly detection and classification tools, suitable for online application, were then applied to the entirety of the Eastern Interconnect Training and Test Datasets. Performance of the event detection and classification tools was evaluated upon receipt of the Test Dataset event logs (i.e., the labels for events contained in the Test Dataset), and promising results were obtained despite several challenges (documented herein) associated with application of supervised or semi-supervised machine learning methods to large-scale, anonymized datasets. Finally, the detection and classification tools were used to detect, classify, and characterize thousands of new events not included in the original event logs provided by the DOE within both the Training and Test Datasets.

24 POWER TRANSMISSION AND DISTRIBUTION↗
NASA NTRSDate not supplied

Foundation AI Models for Science

Foundation Models (FM) are AI models that are designed to replace a task or an application specific model. These FM can be applied to many different downstream applications. These FM are trained using self supervised techniques and can be built on any type of sequence data. The use of self supervised learning removes the hurdle for developing a large labeled dataset for training. Most FM use transformer architecture utilizes the notion of self attention which allows the network to model the influence of distant data points to each other both in space and time. The FM models exhibit emergent properties that are induced from the data. FM can be an important tool for science. The scale of these models results in better performance for different downstream applications and these applications show better accuracy over models built from scratch. FM drastically reduces the cost of entry to build different downstream applications both in time and effort. FM for selected science datasets such as optical satellite data, can accelerate applications ranging from data quality monitoring, feature detection and prediction. FM can make it easier to infuse AI into scientific research by removing the training data bottleneck and increasing the use of science data.

Manil Maskey↗
DOE OSTI2025

RINO: Renormalization Group Invariance with No Labels

A common challenge with supervised machine learning (ML) in high energy physics (HEP) is the reliance on simulations for labeled data, which can often mismodel the underlying collision or detector response. To help mitigate this problem of domain shift, we propose RINO (Renormalization Group Invariance with No Labels), a self-supervised learning approach that can instead pretrain models directly on collision data, learning embeddings invariant to renormalization group flow scales. In this work, we pretrain a transformer-based model on jets originating from quantum chromodynamic (QCD) interactions from the JetClass dataset, emulating real QCD-dominated experimental data, and then finetune on the JetNet dataset -- emulating simulations -- for the task of identifying jets originating from top quark decays. RINO demonstrates improved generalization from the JetNet training data to JetClass data compared to supervised training on JetNet from scratch, demonstrating the potential for RINO pretraining on real collision data followed by fine-tuning on small, high-quality MC datasets, to improve the robustness of ML models in HEP.

Hao, Zichun [Caltech] (ORCID:0000000256244907)↗