Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “massive data sets”

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 19 records
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Rapid detection of rare events from in situ X-ray diffraction data using machine learning

High-energy X-ray diffraction methods can non-destructively map the 3D microstructure and associated attributes of metallic polycrystalline engineering materials in their bulk form. These methods are often combined with external stimuli such as thermo-mechanical loading to take snapshots of the evolving microstructure and attributes over time. However, the extreme data volumes and the high costs of traditional data acquisition and reduction approaches pose a barrier to quickly extracting actionable insights and improving the temporal resolution of these snapshots. This article presents a fully automated technique capable of rapidly detecting the onset of plasticity in high-energy X-ray microscopy data. The technique is computationally faster by at least 50 times than the traditional approaches and works for data sets that are up to nine times sparser than a full data set. This new technique leverages self-supervised image representation learning and clustering to transform massive data sets into compact, semantic-rich representations of visually salient characteristics ( e.g. peak shapes). These characteristics can rapidly indicate anomalous events, such as changes in diffraction peak shapes. It is anticipated that this technique will provide just-in-time actionable information to drive smarter experiments that effectively deploy multi-modal X-ray diffraction methods spanning many decades of length scales.

Zheng, Weijian↗
DOE OSTI2025

Compactly‐Supported Nonstationary Kernels for Computing Exact Gaussian Processes on Big Data

The Gaussian process (GP) is a widely used method for analyzing large-scale data sets, including spatio-temporal measurements of nonlinear processes that are now commonplace in the environmental sciences. Traditional implementations of GPs involve stationary kernels (also termed covariance functions) that limit their flexibility, and exact methods for inference that prevent application to data sets with more than about 10,000 points. Modern approaches to address stationarity assumptions generally fail to accommodate large data sets, while all attempts to address scalability focus on approximating the Gaussian likelihood, which can involve subjectivity and lead to inaccuracies. In this work, we explicitly derive an alternative kernel that can discover and encode both sparsity and nonstationarity. We embed the kernel within a fully Bayesian GP model and leverage high-performance computing resources to enable the analysis of massive data sets. We demonstrate the favorable performance of our novel kernel relative to existing exact and approximate GP methods across a variety of synthetic data examples. Furthermore, we conduct space–time prediction based on more than 1 million measurements of daily maximum temperature and verify that our results outperform state-of-the-art methods in the Earth sciences. More broadly, having access to exact GPs that use ultra-scalable, sparsity-discovering, nonstationary kernels allows GP methods to truly compete with a wide variety of machine learning methods.

Gaussian processes↗
DOE OSTI2025

GLBRC Soil Yearlong Incubation 13C-SIP-Lipidomics

Data package for Lipids represent a dynamic, yet stable pool of microbially-derived soil carbon This data is published under a CC0 license. The authors encourage data reuse and request attribution by referencing the below citations for the data packages and associated manuscript. Please cite as: Rempfert KR, Bell SL, Kasanke CP, Kyle JE, Hofmockel KS. 2025. GLBRC Soil Yearlong Incubation 13C-SIP-Lipidomics. [Data Set] PNNL DataHub. doi: Rempfert KR, Bell SL, Kasanke CP, Kyle JE, Hofmockel KS. 2025. MSV000097435: GLBRC soil yearlong incubation 13C-SIP-Lipidomics [Data Set] MassIVE. doi:10.25345/C57659T3K Rempfert KR, Bell SL, Kasanke CP, Kyle JE, Hofmockel KS. 2025. Lipids represent a dynamic, yet stable pool of microbially-derived soil carbon. In Prep This data package consists of compound-specific 13C SIP-lipidomics data from a yearlong tracer incubation experiment designed to investigate microbial lipid persistence in switchgrass bioenergy crop soils. In order to explore how lipid structure may modulate the persistence of C in soil lipids, we leveraged soils from two sites (Michigan - sandy texture, Wisconsin - silty texture) operated by the U.S. Department of Energy-funded Great Lakes Bioenergy Research Center (GLBRC). These sites had comparable climates, identical management practices, but contrasting soil textures, allowing us to assess the variability of lipid accrual or degradation in soils as well as provide insight regarding the degree to which edaphic properties may regulate the retention of soil lipids. Untargeted lipidomics analyses were performed to identify 13C-labeled lipids in the soil microbiome after long-term incubation. Soils were supplemented with 100 micrograms glucose per gram dry soil (99 atom % 13C or natural abundance for paired control) and incubated; samples were collected two months and one year after glucose addition. Lipid extracts (MPLEx) were analyzed by LC-MS/MS and identified using LIQUID. Calculation of isotopic enrichment of lipids was performed by targeted approach using TarMet to quantify lipid isotopologues and IsoCorrectoR to correct for natural abundance isotopes. Contents: Data package contents reported here are the first version and contain downstream analysis files for the raw LC-MS mass spectrometry files (.mzXML) deposited at the MassIVE database repository under accession MSV000097435 (80 experimental runs; 5.85 GB) | MassIVE DOI: 10.25345/C57659T3K. Support files include the additional data download 'Read Me' file containing data descriptor information. Reported data download contents are structured for compliance with project data sharing guidelines, community standards initiatives, and sponsor stakeholder policies supporting FAIR data principles. Data processing software, analysis tools, and data workflows are listed below corresponding to the host repository long-term location. Available Data Downloads (0.3 GB): "GLBRC soil yearlong incubation 13C-SIP-Lipidomics_readme.txt" - 'Read Me' data package content file (txt) "GLBRC_DataPackage_analysis files" - Data processing files (Rmd) and saved intermediate data processing outputs (rds, csv, xlsx) "GLBRC_13C_lipidomics_dataset.xlsx" - processed data in tabular format (xlsx) Linked Software: LIQUID LC-MS Analysis Software | 10.5281/zenodo.6459462 Lipid Mini-On Software Tools | 10.5281/zenodo.1492803 pmartR Omics Statistical Software | 10.5281/zenodo.6108667 xcms (v4.3.3) TarMet (v1.1.1) IsoCorrectoR (1.24.0) Funding Acknowledgments: This research was supported by an Early Career Research Program award funded by the U.S. Department of Energy, Office of Science, Office of Biological and Environmental Research (OBER) Genomic Science program under FWP 68292, FWP 07880 and EMSL Exploratory Research Project 51095. A portion of this work was performed in the William R. Wiley Environmental Molecular Sciences Laboratory, a national scientific user facility sponsored by OBER and located at Pacific Northwest National Laboratory (PNNL). PNNL is a multi-program national laboratory operated by Battelle for the DOE under Contract DE-AC05-76RLO1830.

Rempfert, Kaitlin R [Pacific Northwest National La↗
DOE OSTI2025

Survey-wide asteroid discovery with a high-performance computing enabled non-linear digital tracking framework

Modern astronomical surveys detect asteroids by linking together their appearances across multiple images taken over time. This approach faces limitations in detecting faint asteroids and handling the computational complexity of trajectory linking. Here, we present a novel method that adapts “digital tracking” – traditionally used for short-term linear asteroid motion across images – to work with large-scale synoptic surveys such as the Vera Rubin Observatory Legacy Survey of Space and Time (Rubin/LSST). Our approach combines hundreds of sparse observations of individual asteroids across their non-linear orbital paths to enhance detection sensitivity by several magnitudes. To address the computational challenges of processing massive data sets and dense orbital phase spaces, we developed a specialized high-performance computing architecture. We demonstrate the effectiveness of our method through experiments that take advantage of the extensive computational resources at Lawrence Livermore National Laboratory. This work enables the detection of significantly fainter asteroids in existing and future survey data, potentially increasing the observable asteroid population by orders of magnitude across different orbital families, from near-Earth objects (NEOs) to Kuiper belt objects (KBOs).

Asteroid discovery↗
DOE OSTI2024

UMap: An application-oriented user level memory mapping library

Exploiting the prominent role of complex memories in exascale node architecture, the UMap page fault handler offers new capabilities to access large memory-mapped data sets directly. UMap provides flexible configuration options to customize page handling to each application, including analysis of massive observational and simulation data sets. The high-performance design features I/O decoupling, dynamic load balancing, and application-level controls. Page faults triggered by application threads and processes accessing data mapped to a UMapp’ed region are handled via the Linux userfaultfd protocol, an asynchronous message-oriented kernel-user communication mechanism that avoids the context switch penalty of traditional signal fault handlers. UMap is fully open source. In this paper, we give an overview of the UMap library architecture, its extensible plugin architecture, and the use/performance of UMap in emerging heterogeneous memory hierarchies such as near-node Non-volatile Memory (NVM) and network attached memories. We highlight new capabilities in two pagefault management plugins, the NetworkStore and SparseStore. We demonstrate the integration between UMap and multiple ECP products including Caliper, Metall, ZFP, Mochi, and Ripples.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2026

Artificial Intelligence Research at Fermilab

The unique challenges of high-energy physics research present opportunities for advancing AI technologies. From the principles of fundamental physics underlying massive and rich data sets to building and operating some of the world’s most complex detector and accelerator systems, the technologies we are developing have potential connections to a broad domain of cutting-edge AI research.

Fermilab, Fermilab↗
DOE OSTI2025

Elucidating texture and grain morphology contributions to the micromechanical response of additively manufactured Inconel 625

Microstructural variation of additively manufactured (AM) metal components in comparison to wrought counterparts makes certification for critical applications a challenge. Microscale simulations leveraging modern computational tools may be used to supplement testing of AM microstructures, thus accelerating certification by reducing the number of experiments needed. However, as micromechanical response is closely tied to critical properties like fatigue-life and fracture, utilization of these simulations with macroscale experimental data alone is insufficient. One means to attain microscale experimental data is in situ diffraction data collected from synchrotron X-ray sources. In this work, such data were collected during in situ compression of AM Inconel 625 superalloy. Interpretation of experimental results was assisted by massive (8M element) complementary micromechanical simulations performed on sets of virtual microstructures generated using cellular automata. Together, micromechanical data from diffraction experiments and simulations were used to probe the effects of textured “track” microstructures generated during laser powder bed fusion and directional strength-to-stiffness on micromechanical response. Though fiber-averaged directional strength-to-stiffness ratios were expected to dominate given the high elastic anisotropy of the material, the combination of small variations in texture and specific grain configurations unique to AM microstructures lead to significant variability in micromechanical response after yield. The findings emphasize the importance of high-fidelity microstructural representation that captures key texture components and AM-specific morphology for property prediction of AM metals.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Does the Z boson have a lighter cousin?

In the quenched electroweak theory on the lattice I construct a set of physical states which overlap the physical photon and Z boson states. This is done by employing eigenstates of the covariant lattice Laplacian, in addition to the Higgs and lattice link variables, to construct gauge invariant vector boson creation operators. Diagonalizing the transfer matrix in the subspace of Hilbert space spanned by this set yields a massless photon and massive Z particle, as expected. But in the numerical data there is evidence for more vector bosons in the spectrum, albeit with considerable uncertainty in their masses, with the lowest finite mass particle in the range of 3–4 GeV. Published by the American Physical Society 2025

Greensite, Jeff (ORCID:0000000317209436)↗
DOE OSTI2025

A Parametric, Data-Driven, Non-Intrusive Reduced-Order Model Framework for Crystal Plasticity Simulations of Voids

The influence of the internal structure at micrometer length scales on the deformation of polycrystalline materials can be effectively captured using crystal plasticity finite element methods (CPFEM). However, the complexity and nonlinearity of the deformation equations CPFEM solves demand significant computational power and resources to achieve accurate predictions, limiting its broader application. To address this challenge, we have identified a reduced-order representation of the complex data in order to establish a computationally efficient reduced-order models (ROM) and drastically reduce the computational expense of CPFEM. Specifically, in this work, we developed a parametric, data-driven, and non-intrusive ROM framework for CPFEM using proper orthogonal decomposition (POD) and sparse variational Gaussian process (SVGP) regression for single-crystal microstructures under tensile loading conditions. The developed protocol enables one to compress field into a latent/low-dimensional space described by principal component analysis (PCA) via the singular value decomposition (SVD) algorithm. As a result, the high-dimensional data are reduced to a significantly smaller amount of dimensions with POD bases and POD coefficients. Furthermore, we deployed an ensemble of SVGPs—extended from the classical Gaussian process (GP) regression for scalability and handling big data—in a massively parallel manner to train and predict latent POD coefficients using known POD bases from a set of previously obtained simulations results. Lastly, using the predicted POD coefficients, we reconstructed the full-field results and showed reasonable agreement compared with the true values obtained from running CPFEM. The developed framework is validated with a set of CPFEM simulations of a single embedded void in single-crystal aluminum alloy. While the framework is broadly applicable, this work specifically focuses on single-crystal microstructures, a single load case (e.g., tensile), and a specific void geometry (spherical).

Anisotropy↗
DOE OSTI2024

A Million Person Study Innovation: Evaluating Cognitive Impairment and other Morbidity Outcomes from Chronic Radiation Exposure Through Linkages with the Centers for Medicaid and Medicare Services Assessment and Claims Data

Here, the study of One Million U.S. Radiation Workers and Veterans, the Million Person Study (MPS), examines the health consequences, both cancer and non-cancer, of exposure to ionizing radiation received gradually over time. Recently the MPS has focused on mortality patterns from neurological and behavioral conditions, e.g., Parkinson's disease, Alzheimer's disease, dementia, and motor neuron disease such as amyotrophic lateral sclerosis. A fuller picture of radiation-related late effects comes from studying both mortality and the occurrence (incidence) of conditions not leading to death. Accordingly, the MPS is identifying neurocognitive diagnoses from fee-for-service insurance claims from the Centers for Medicare and Medicaid Services (CMS), among Medicare beneficiaries beginning in 1999 (the earliest date claims data are available). Linkages to date have identified ∼540,000 workers with available health information. Such linkages provide individual information on important co-factor and confounding variables such as smoking, alcohol consumption, blood pressure, obesity, diabetes and many other health and demographic characteristics. The total person-level set of time-dependent variables, outcomes, organ-specific dose measures, co-factors, and demographics will be massive and much too large to be evaluated with standard software. Thus, development of specialized open-source software designed for large datasets (Colossus) is nearly complete. The wealth of information available from CMS claims data, coupled with individual dose reconstructions, will thus greatly enhance the quality and precision of health evaluations for this new field of low-dose radiation and neurocognitive effects.

Dauer, Lawrence T.↗
DOE OSTI2025

Scalable edge clustering of dynamic graphs via weighted line graphs

Timestamped relational datasets consisting of records (or connections) between pairs of entities are ubiquitous in network science. For applications like peer-to-peer communication, email, various social network interactions, and computer network security, it is useful to organize these records into groups based on how and when they are occurring. Weighted line graphs offer a natural way to model how records are related in such datasets but for large real-world graph topologies, building and utilizing the line graph is prohibitively expensive. Here, we present the framework to cluster the edges of a dynamic graph via the associated line graph that contains two major contributions. The first is a method to work with the line graph implicitly and the second is a distributed scale implementation of an agglomerative hierarchical graph clustering algorithm. We outline a novel hierarchical dynamic graph edge clustering approach that efficiently breaks massive relational datasets into small sets of edges containing events at various timescales. This is in stark contrast to traditional graph clustering algorithms that prioritize highly connected (clique-like) community structures. Our approach relies on constructing a sufficient subgraph of a weighted line graph and applying a hierarchical agglomerative clustering. This approach is related to scalable techniques from spatial clustering, nonlinear-dimension reduction, topological data analysis, and draws particular inspiration from HDBSCAN. As an edge clustering, this method yields an overlapping node clustering. Our algorithm is parallelizable and we demonstrate efficient clustering of a billion-scale, real-world dynamic graph into small edge sets that correlate in topology and time. The entire clustering process for a graph with tens of billions of edges takes just a few minutes of run time on 256 nodes of a distributed compute environment. We argue how the output of the edge clustering is useful for a multitude of data visualization and powerful machine learning tasks, both involving the original massive dynamic graph data and metadata associated with the nodes and edges. Finally, we describe how this approach can be extended to dynamic hypergraphs and dynamic graphs/hypergraphs with unstructured data living on vertices and edges.

Data Analysis↗
DOE OSTI2026

Sensitivity to low-mass WIMPs with an improved liquid argon ionization response model within the DarkSide program

Dark matter detection experiments using liquid argon rely on a precise characterization of the ionization response to nuclear recoils, especially in the keV energy range relevant for light dark matter interactions. In this work, we present a comprehensive analysis that combines new measurements from the ReD setup, part of the DarkSide experimental program, with calibration data from DarkSide-50, as well as results from the ARIS and SCENE experiments. These combined datasets enable improved constraints on atomic screening effects in the modeling of the ionization response of liquid argon to nuclear recoils. The analysis is performed within the Thomas-Imel recombination framework adopted in previous DarkSide studies, and is here further constrained by the inclusion of ReD data, which allow the screening function to be determined from calibration measurements. By including the updated ionization model into the DarkSide-50 analysis framework, we obtain stronger exclusion limits on low-mass weakly interacting massive particle (WIMP) interactions, setting new world-leading constraints in the 1 – 3 GeV / c 2 WIMP mass range. Finally, we recast the sensitivity projections for the upcoming DarkSide-20k detector, demonstrating a significantly enhanced discovery potential for low-mass dark matter candidates.

Acerbi, F. [Fond. Bruno Kessler, Trento]↗
DOE OSTI2024

Utility-Scale Solar, 2024 Edition: Empirical Trends in Deployment, Technology, Cost, Performance, PPA Pricing, and Value in the United States [Slides]

Berkeley Lab’s “Utility-Scale Solar, 2024 Edition” presents analysis of empirical plant-level data from the U.S. fleet of ground-mounted photovoltaic (PV), PV+battery, and concentrating solar-thermal power (CSP) plants with capacities exceeding 5 MWAC (PV plants of 5 MWAC or less, including residential rooftop systems, are covered separately in Berkeley Lab’s companion annual report, Tracking the Sun). Key findings from this year’s report include: -18.5 GWAC of new utility-scale PV capacity came online in 2023, bringing cumulative installed capacity to more than 80.2 GWAC across 47 states. Installed costs continued to fall in 2023. Relative to 2022, capacity-weighted averages decreased by 8% to -$\$1.43$/WAC (or $\$1.08$/WDC). Costs, based on a 7.1 GWAC sample of 76 plants completed in 2023, have fallen by 75% (averaging 10% annually) since 2010. Plant-level capacity factors vary widely, from 6% to 36% (on an AC basis), with a sample median of 24%. -Levelized cost of energy (LCOE) of new 2023 projects increased slightly to $\$46$/MWh prior to the application of tax credits but continued to fall to $\$31$/MWh when accounting for federal incentives. PPA prices have largely followed the decline in solar’s LCOE over time, but newly signed longer-term PPA prices have increased since 2021, to an average of $\$35$/MWh (levelized, in 2023 dollars). -Solar’s average energy and capacity value (i.e., ability to offset costs of other power generation sources) across the U.S. was $\$45$/MWh in 2023. Solar’s average market value was lowest in CAISO ($\$27$/MWh), the market with the greatest solar generation share, and highest in ERCOT ($\$67$/MWh). -Newer solar projects had greater market value in 2023 than their generation costs, yielding $\$1.1$ billion in benefits. Projects built in 2022 delivered on average $\$15$/MWh more market value than their costs in 2023. -Solar’s combined value from wholesale electricity markets, public health and climate damage reduction were greater than generation costs and incentives, yielding $\$13.7$ billion in net benefits in 2023. We estimate U.S. health benefits of $\$24$/MWh and reduced global climate damages of $\$101$/MWh. -Adding battery storage is one way to increase the value of solar. Deployment of 52 new PV+battery hybrid plants set a record with 5.3 GW installed in 2023. Our public data file tracks metadata and PPA prices from more than 100 PV+battery hybrid projects that are already online or that have secured offtake arrangements. -Looking ahead, a massive pipeline of at least 1,085 GW of solar capacity dominates the nation’s interconnection queues at the end of 2023. Nearly 571 GW, or 53%, of that total was paired with a battery – in CAISO it was a staggering 98%. Historically only 10% of the requested solar capacity is built. -For more information, and to explore related interactive data visualizations, go to utilityscalesolar.lbl.gov.

14 SOLAR ENERGY↗
DOE OSTI2024

Linking transients to their host galaxies – II. A comparison of host galaxy properties and rate dependencies across supernova types

ABSTRACT We use the latest data set of supernova (SN) host galaxies to investigate how the host properties – stellar mass, star formation rate (SFR), metallicity, absolute magnitude, and colour–differ across SN types, with redshift-driven selection effects controlled. SN Ib and Ic host galaxies, on average, are more massive, metal-rich, and redder than SN II hosts. For subtypes, SN Ibn and Ic-BL have bluer hosts than their normal SN Ib and Ic siblings; SN IIb has consistent host properties with SN Ib, while hosts of SN IIn are more metal-rich than those of SN II. Hydrogen-deficient superluminous SNe feature bluer and lower luminosity hosts than most subtypes of core-collapse supernova (CC SN). Assuming simple proportionality of CC SN rates and host SFRs does not recover the observed mean host properties; either a population of long-lived progenitors or a metallicity-dependent SN production efficiency better reproduces the observed host properties. Assuming the latter case, the rates of SN II are insensitive to host metallicity, but the rates of SN Ib and Ic are substantially enhanced in metal-rich hosts by a factor of $\sim 10$ per dex increase in metallicity. Hosts of SN Ia are diverse in their observed properties; subtypes including SN Ia-91T, Ia-02cx, and Ia-CSM prefer star-forming hosts, while subtypes like SN Ia-91bg and Ca-rich prefer quiescent hosts. The rates of SN Ia-91T, Ia-02cx, and Ia-CSM are closely dependent on, or even proportional to, their host SFRs, indicating relatively short-lived progenitors. Conversely, the rates of SN Ia-91bg and Ca-rich transients are proportional to the total stellar mass, favouring long-lived progenitors.

Qin, Yu-Jing (ORCID:0000000336586026)↗
DOE OSTI2025

Comparing Compressed and Full-Modeling analyses with FOLPS: implications for DESI 2024 and beyond

The Dark Energy Spectroscopic Instrument (DESI) will provide unprecedented information about the large-scale structure of our Universe. In this work, we study the robustness of the theoretical modelling of the power spectrum of F OLPS , a novel effective field theory-based package for evaluating the redshift space power spectrum in the presence of massive neutrinos. We perform this validation by fitting the AbacusSummit high-accuracy N -body simulations for Luminous Red Galaxies, Emission Line Galaxies and Quasar tracers, calibrated to describe DESI observations. We quantify the potential systematic error budget of F OLPS finding that the modelling errors are fully sub-dominant for the DESI statistical precision within the studied range of scales. Additionally, we study two complementary approaches to fit and analyse the power spectrum data, one based on direct Full-Modelling fits and the other on the ShapeFit compression variables, both resulting in very good agreement in precision and accuracy. In each of these approaches, we study a set of potential systematic errors induced by several assumptions, such as the choice of template cosmology, the effect of prior choice in the nuisance parameters of the model, or the range of scales used in the analysis. Furthermore, we show how opening up the parameter space beyond the vanilla ΛCDM model affects the DESI observables. These studies include the addition of massive neutrinos, spatial curvature, and dark energy equation of state. We also examine how relaxing the usual Cosmic Microwave Background and Big Bang Nucleosynthesis priors on the primordial spectral index and the baryonic matter abundance, respectively, impacts the inference on the rest of the parameters of interest. This paper pathways towards performing a robust and reliable analysis of the shape of the power spectrum of DESI galaxy and quasar clustering using F OLPS .

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2024

Efficient distributed continual learning for steering experiments in real-time

Deep learning has emerged as a powerful method for extracting valuable information from large volumes of data. However, when new training data arrives continuously (i.e., is not fully available from the beginning), incremental training suffers from catastrophic forgetting (i.e., new patterns are reinforced at the expense of previously acquired knowledge). Training from scratch each time new training data becomes available would result in extremely long training times and massive data accumulation. Rehearsal-based continual learning has shown promise for addressing the catastrophic forgetting challenge, but research to date has not addressed performance and scalability. To fill this gap, we propose an approach based on a distributed rehearsal buffer that efficiently complements data-parallel training on multiple GPUs to achieve high accuracy, short runtime, and scalability. It leverages a set of buffers (local to each GPU) and uses several asynchronous techniques for updating these local buffers in an embarrassingly parallel fashion, all while handling the communication overheads necessary to augment input minibatches using unbiased, global sampling. We further propose a generalization of rehearsal buffers to support both classification and generative learning tasks, as well as more advanced rehearsal strategies (notably Dark Experience Replay, leveraging knowledge distillation). We illustrate this approach with a real-life HPC streaming application from the domain of ptychographic image reconstruction. Furthermore, we run extensive experiments on up to 128 GPUs of the ThetaGPU supercomputer to compare our approach with baselines representative of training-from-scratch (the upper bound in terms of accuracy) and incremental training (the lower bound). Results show that rehearsal-based continual learning achieves a top-5 validation accuracy close to the upper bound, while simultaneously exhibiting a runtime close to the lower bound.

Asynchronous data management↗
DOE OSTI2024

Search for a scalar or pseudoscalar dilepton resonance produced in association with a massive vector boson or top quark-antiquark pair in multilepton events at s = 13 TeV

A search for beyond the standard model spin-0 bosons, ϕ , that decay into pairs of electrons, muons, or tau leptons is presented. The search targets the associated production of such bosons with a W or Z gauge boson, or a top quark-antiquark pair, and uses events with three or four charged leptons, including hadronically decaying tau leptons. The proton-proton collision data set used in the analysis was collected at the LHC from 2016 to 2018 at a center-of-mass energy of 13 TeV, and corresponds to an integrated luminosity of 138 fb - 1 . The observations are consistent with the predictions from standard model processes. Upper limits are placed on the product of cross sections and branching fractions of such new particles over the mass range of 15 to 350 GeV with scalar, pseudoscalar, or Higgs-boson-like couplings, as well as on the product of coupling parameters and branching fractions. Several model-dependent exclusion limits are also presented. For a Higgs-boson-like ϕ model, limits are set on the mixing angle of the Higgs boson with the ϕ boson. For the associated production of a ϕ boson with a top quark-antiquark pair, limits are set on the coupling to top quarks. Finally, limits are set for the first time on a fermiophilic dilaton-like model with scalar couplings and a fermiophilic axion-like model with pseudoscalar couplings.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2025

Hardware acceleration for HPS algorithms in two and three dimensions

We provide a flexible, open-source framework for hardware acceleration, namely massively-parallel execution on general-purpose graphics processing units (GPUs), applied to the hierarchical Poincaré–Steklov (HPS) family of algorithms for building fast direct solvers for linear elliptic partial differential equations. To take full advantage of the power of hardware acceleration, we propose two variants of HPS algorithms to improve performance on two- and three-dimensional problems. In the two-dimensional setting, we introduce a novel recomputation strategy that minimizes costly data transfers to and from the GPU; in three dimensions, we modify and extend the adaptive discretization technique of Geldermans and Gillman [1] to greatly reduce peak memory usage. We provide an open-source implementation of these methods written in JAX, a high-level accelerated linear algebra package, which allows for the first integration of a high-order fast direct solver with automatic differentiation tools. We conclude with extensive numerical examples showing our methods are fast and accurate on two- and three-dimensional problems.

Fast direct solvers↗