Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Computational pipelines”

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 451 records · Page 25
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

HEPTAPOD: Orchestrating High Energy Physics Workflows Towards Autonomous Agency

Many workflows in high-energy-physics (HEP) stand to benefit from recent advances in transformer-based large language models (LLMs). While early applications of LLMs focused on text generation and code completion, modern LLMs now support orchestrated agency: the coordinated execution of complex, multi-step tasks through tool use, structured context, and iterative reasoning. We introduce the HEP Toolkit for Agentic Planning, Orchestration, and Deployment (HEPTAPOD), an orchestration framework designed to bring this emerging paradigm to HEP pipelines. The framework enables LLMs to interface with domain-specific tools, construct and manage simulation workflows, and assist in common utility and data analysis tasks through schema-validated operations and run-card-driven configuration. To demonstrate these capabilities, we consider a representative Beyond the Standard Model (BSM) Monte Carlo validation pipeline that spans model generation, event simulation, and downstream analysis within a unified, reproducible workflow. HEPTAPOD provides a structured and auditable layer between human researchers, LLMs, and computational infrastructure, establishing a foundation for transparent, human-in-the-loop systems.

Menzo, Tony [Alabama U.; Fermilab] (ORCID:00000002↗
DOE OSTI2024

Navigating Exascale Operational Data Analytics: From Inundation to Insight

In this paper, we address the challenges in achieving sustainable data-driven efficiency by providing a detailed exploration of the end-to-end operational data analytics (ODA) framework that evolved through two generations of supercomputer systems at the Oak Ridge Leadership Computing Facility (OLCF). This framework addresses large data streams ingested from heavily instrumented HPC environment that accumulates multi-terabytes per day. We outline the multifaceted data life cycle across HPC procurement, operations, and research & development, identifying key obstacles and design decisions that shape effective strategies in building and supporting data pipelines end-to-end. By sharing key insights and lessons learned from our experience, we offer recommendations for the HPC community on enabling sustainable operational data analytics and beyond. Our contributions aim to bridge the gap between potential and real benefits of operational data, guiding future efforts towards integrated and sustainable operational intelligence in high-performance computing environments.

Shin, Woong↗
DOE OSTI2022

Inductive predictions of hydrologic events using a Long Short-Term Memory network and the Soil and Water Assessment Tool

We present machine learning methods to predict hydrologic features such as streamflow and soil moisture from spatially and temporally varying hydrological and meteorological data. Here, we used a temporal reduction technique to reduce computation and memory requirements and trained a Long Short-Term Memory (LSTM) network to predict soil moisture and streamflow over multiple watersheds. We show LSTM networks can be trained in a fraction of the time required by complex process-based and attention-based models such as Soil and Water Assessment Tool (SWAT) and GeoMAN without sacrificing accuracy. We also demonstrate that outside data - sourced from a watershed other than the target - can be used to train LSTM to comparable or even superior prediction accuracy. The success of LSTM in such spatially-inductive settings shows hydrologic features can be predicted with minimal prior knowledge of the watershed in question. Finally, we make all methodologies of this work publicly available as an end-to-end software pipeline that facilitates rapid prototyping of hydrologic learners.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Outlook for artificial intelligence and machine learning at the NSLS-II

Abstract We describe the current and future plans for using artificial intelligence and machine learning (AI/ML) methods at the National Synchrotron Light Source II (NSLS-II), a scientific user facility at the Brookhaven National Laboratory. We discuss the opportunity for using the AI/ML tools and techniques developed in the data and computational science areas to greatly improve the scientific output of large scale experimental user facilities. We describe our current and future plans in areas including from detecting and recovering from faults, optimizing the source and instrument configurations, streamlining the pipeline from measurement to insight, through data acquisition, processing, analysis. The overall strategy and direction of the NSLS-II facility in relation to AI/ML is presented.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

A Length Adaptive Algorithm-Hardware Co-design of Transformer on FPGA Through Sparse Attention and Dynamic Pipelining

Transformers are considered one of the most important deep learning models since 2018, in part because it establishes state-of-the-art (SOTA) records and could potentially replace existing Deep Neural Networks (DNNs). Despite the remarkable triumphs, the prolonged turnaround time of Transformer models is a widely recognized roadblock. The variety of sequence lengths imposes additional computing overhead where inputs need to be zero-padded to the maximum sentence length in the batch to accommodate the parallel computing platforms. This paper targets the field-programmable gate array (FPGA) and proposes a coherent sequence length adaptive algorithm–hardware co-design for Transformer acceleration. Particularly, we develop a hardware-friendly sparse attention operator and a length-aware hardware resource scheduling algorithm. The proposed sparse attention operator brings the complexity of attention-based models down to linear complexity and alleviates the off-chip memory traffic. The proposed length-aware resource hardware scheduling algorithm dynamically allocates the hardware resources to fill up the pipeline slots and eliminates bubbles for NLP tasks. Experiments show that our design has very small accuracy loss and has 80.2 × and 2.6 × speedup compared to CPU and GPU implementation, and 4 × higher energy efficiency than state-of-the-art GPU accelerator optimized via CUBLAS GEMM.

Peng, Hongwu↗
DOE OSTI2022

Cosmic shear in harmonic space from the Dark Energy Survey Year 1 Data: compatibility with configuration space results

ABSTRACT We perform a cosmic shear analysis in harmonic space using the first year of data collected by the Dark Energy Survey (DES-Y1). We measure the cosmic weak lensing shear power spectra using the metacalibration catalogue and perform a likelihood analysis within the framework of CosmoSIS. We set scale cuts based on baryonic effects contamination and model redshift and shear calibration uncertainties as well as intrinsic alignments. We adopt as fiducial covariance matrix an analytical computation accounting for the mask geometry in the Gaussian term, including non-Gaussian contributions. A suite of 1200 lognormal simulations is used to validate the harmonic space pipeline and the covariance matrix. We perform a series of stress tests to gauge the robustness of the harmonic space analysis. Finally, we use the DES-Y1 pipeline in configuration space to perform a similar likelihood analysis and compare both results, demonstrating their compatibility in estimating the cosmological parameters S8, σ8, and Ωm. We use the DES-Y1 metacalibration shape catalogue, with photometric redshifts estimates in the range of 0.2−1.3, divided in four tomographic bins finding σ8(Ωm/0.3)0.5 = 0.766 ± 0.033 at 68 per cent CL. The methods implemented and validated in this paper will allow us to perform a consistent harmonic space analysis in the upcoming DES data.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2023

Why is the winner the best?

International benchmarking competitions have become fundamental for the comparative performance assessment of image analysis methods. However, little attention has been given to investigating what can be learnt from these competitions. Do they really generate scientific progress? What are common and successful participation strategies? What makes a solution superior to a competing method? To address this gap in the literature, we performed a multi- center study with all 80 competitions that were conducted in the scope of IEEE ISBI 2021 and MICCAI 2021. Statistical analyses performed based on comprehensive descriptions of the submitted algorithms linked to their rank as well as the underlying participation strategies revealed common characteristics of winning solutions. These typically include the use of multi-task learning (63%) and/or multi-stage pipelines (61%), and a focus on augmentation (100%), im- age preprocessing (97%), data curation (79%), and post- processing (66%). The “typical” lead of a winning team is a computer scientist with a doctoral degree, five years of experience in biomedical image analysis, and four years of experience in deep learning. Two core general development strategies stood out for highly-ranked teams: the reflection of the metrics in the method design and the focus on analyzing and handling failure cases. According to the organizers, 43% of the winning algorithms exceeded the state of the art but only 11% completely solved the respective domain problem. The insights of our study could help researchers (1) improve algorithm development strategies when approaching new problems, and (2) focus on open research questions revealed by this work.

Eisenmann, Matthias↗
NASA NTRS1993

Feed-forward volume rendering algorithm for moderately parallel MIMD machines

Algorithms for direct volume rendering on parallel and vector processors are investigated. Volumes are transformed efficiently on parallel processors by dividing the data into slices and beams of voxels. Equal sized sets of slices along one axis are distributed to processors. Parallelism is achieved at two levels. Because each slice can be transformed independently of others, processors transform their assigned slices with no communication, thus providing maximum possible parallelism at the first level. Within each slice, consecutive beams are incrementally transformed using coherency in the transformation computation. Also, coherency across slices can be exploited to further enhance performance. This coherency yields the second level of parallelism through the use of the vector processing or pipelining. Other ongoing efforts include investigations into image reconstruction techniques, load balancing strategies, and improving performance.

Yagel, Roni↗
NASA NTRS2010

The Kepler Science Operations Center Pipeline Framework Extensions

The Kepler Science Operations Center (SOC) is responsible for several aspects of the Kepler Mission, including managing targets, generating on-board data compression tables, monitoring photometer health and status, processing the science data, and exporting the pipeline products to the mission archive. We describe how the generic pipeline framework software developed for Kepler is extended to achieve these goals, including pipeline configurations for processing science data and other support roles, and custom unit of work generators that control how the Kepler data are partitioned and distributed across the computing cluster. We describe the interface between the Java software that manages the retrieval and storage of the data for a given unit of work and the MATLAB algorithms that process these data. The data for each unit of work are packaged into a single file that contains everything needed by the science algorithms, allowing these files to be used to debug and evolve the algorithms offline.

Klaus, Todd C.↗
NASA NTRS2016

Erasing the Milky Way: New Cleaning Technique Applied to GBT Intensity Mapping Data

We present the first application of a new foreground removal pipeline to the current leading HI intensity mapping dataset, obtained by the Green Bank Telescope (GBT). We study the 15- and 1-h field data of the GBT observations previously presented in Masui et al. (2013) and Switzer et al. (2013), covering about 41 square degrees at 0.6 less than z is less than 1.0, for which cross-correlations may be measured with the galaxy distribution of the WiggleZ Dark Energy Survey. In the presented pipeline, we subtract the Galactic foreground continuum and the point source contamination using an independent component analysis technique (fastica), and develop a Fourier-based optimal estimator to compute the temperature power spectrum of the intensity maps and cross-correlation with the galaxy survey data. We show that fastica is a reliable tool to subtract diffuse and point-source emission through the non-Gaussian nature of their probability distributions. The temperature power spectra of the intensity maps is dominated by instrumental noise on small scales which fastica, as a conservative sub-traction technique of non-Gaussian signals, can not mitigate. However, we determine similar GBT-WiggleZ cross-correlation measurements to those obtained by the Singular Value Decomposition (SVD) method, and confirm that foreground subtraction with fastica is robust against 21cm signal loss, as seen by the converged amplitude of these cross-correlation measurements. We conclude that SVD and fastica are complementary methods to investigate the foregrounds and noise systematics present in intensity mapping datasets.

Wolz, L.↗
DOE OSTI2025

Enhanced polymorph metastability drives glycine nucleation in aqueous salt solutions

Crystal nucleation from aqueous solutions influences countless geological, biochemical, astrophysical, environmental, and materials science–related phenomena, including ice formation, the manufacturing of active pharmaceutical ingredients, development of diseases such as Alzheimer’s and the origin of life itself. Understanding and controlling nucleation is essential for designing materials with specific properties, developing strategies to inhibit or promote crystallization in various contexts and preventing pathological aggregation in neurodegenerative diseases. Similar to the protein structure prediction problem—where a single amino acid sequence can in theory adopt one most stable conformation but in practice may sample multiple competing conformations—crystal nucleation faces a parallel challenge: the same chemical species can form diverse polymorphs under different environmental conditions (e.g., temperature, pressure, solvent). Each polymorph presents its own set of physical and chemical properties, highlighting the importance of understanding and controlling polymorph selection in fields ranging from pharmaceuticals to materials design. Despite advances in experimental and computational methods for studying phase transitions and polymorph stability, nucleation remains challenging due to its nanoscale nature. Furthermore, in practical settings, salts and impurities can further influence crystal nucleation in diverse contexts, from scaling in pipelines and desalination plants to the durability of concrete and the efficiency of battery materials. This can lead to the formation of polymorphs that may differ from the most stable phase in pure solutions. Or, even though the final structure might appear same irrespective of whether the environment contained impurities or not, the mechanism through which it was formed might be completely different and not intuitive.

Wang, Ruiyu [University of Maryland, College Park,↗
DOE OSTI2024

TomoPyUI : a user-friendly tool for rapid tomography alignment and reconstruction

The management and processing of synchrotron and neutron computed tomography data can be a complex, labor-intensive and unstructured process. Users devote substantial time to both manually processing their data ( i.e. organizing data/metadata, applying image filters etc. ) and waiting for the computation of iterative alignment and reconstruction algorithms to finish. In this work, we present a solution to these problems: TomoPyUI , a user interface for the well known tomography data processing package TomoPy . This highly visual Python software package guides the user through the tomography processing pipeline from data import, preprocessing, alignment and finally to 3D volume reconstruction. The TomoPyUI systematic intermediate data and metadata storage system improves organization, and the inspection and manipulation tools (built within the application) help to avoid interrupted workflows. Notably, TomoPyUI operates entirely within a Jupyter environment. Herein, we provide a summary of these key features of TomoPyUI , along with an overview of the tomography processing pipeline, a discussion of the landscape of existing tomography processing software and the purpose of TomoPyUI , and a demonstration of its capabilities for real tomography data collected at SSRL beamline 6-2c.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2021

Understanding the Design Space of Sparse/Dense Multiphase Dataflows for Mapping Graph Neural Networks on Spatial Accelerators

Graph Neural Networks (GNNs) have garnered a lot of recent interest because of their success in learning representations from graph-structured data across several critical applications in cloud and HPC. Owing to their unique compute and memory characteristics that come from an interplay between dense and sparse phases of computations, the emergence of reconfigurable dataflow (aka spatial) accelerators offers promise for acceleration by mapping optimized dataflows (i.e., computation order and parallelism) for both phases. The goal of this work is to characterize and understand the design-space of dataflow choices for running GNNs on spatial accelerators in order for the compilers to optimize the dataflow based on the workload. Specifically, we propose a taxonomy to describe all possible choices for mapping the dense and sparse phases of GNNs spatially and temporally over a spatial accelerator, capturing both the intra-phase dataflow and the inter-phase (pipelined) dataflow. Using this taxonomy, we do deep-dives into the cost and benefits of several dataflows and perform case studies on implications of hardware parameters for dataflows and value of flexibility to support pipelined execution.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Full forward model of galaxy clustering statistics with AbacusSummit light cones

ABSTRACT Novel summary statistics beyond the standard 2-point correlation function (2PCF) are necessary to capture the full astrophysical and cosmological information from the small-scale (r < 30h−1Mpc) galaxy clustering. However, the analysis of beyond-2PCF statistics on small scales is challenging because we lack the appropriate treatment of observational systematics for arbitrary summary statistics of the galaxy field. In this paper, we develop a full forward modelling pipeline for a wide range of summary statistics using the large high-fidelity AbacusSummit light cones that account for many systematic effects as well as remain flexible and computationally efficient to enable posterior sampling. We apply our forward model approach to a fully realistic mock galaxy catalog and demonstrate that we can recover unbiased constraints on the underlying galaxy–halo connection model using two separate summary statistics: the standard 2PCF and the novel k-th nearest neighbour (kNN) statistics, which are sensitive to correlation functions of all orders. We will demonstrate its strong constraining power on extended galaxy–halo connection models and cosmology in follow up papers. We expect this to become a powerful approach when applying to upcoming surveys such as DESI where we can leverage a multitude of summary statistics across a wide redshift range to maximally extract information from the non-linear scales.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2025

A knowledge-informed large language model framework for U.S. nuclear power plant shutdown initiating event classification for probabilistic risk assessment

Identifying and classifying shutdown initiating events (SDIEs) is critical for developing shutdown probabilistic risk assessment for nuclear power plants. Existing computational approaches cannot achieve satisfactory performance due to the challenges of unavailable large, labeled datasets, imbalanced event types, and label noise. To address these challenges, we propose a hybrid pipeline that integrates a knowledge-informed machine learning model to prescreen non-SDIEs and a large language model (LLM) to classify SDIEs into four types. In the prescreening stage, we proposed a set of 44 SDIE text patterns that consist of the most salient keywords and phrases from six SDIE types. Text vectorization based on the SDIE patterns generates feature vectors that are highly separable by using a simple binary classifier. The second stage builds Bidirectional Encoder Representations from Transformers (BERT)-based LLM, which learns generic English language representations from self-supervised pretraining on a large dataset and adapts to SDIE classification by fine-tuning it on an SDIE dataset. The proposed approaches are evaluated on a dataset with 10,928 events using precision, recall ratio, F 1 score, and average accuracy. In conclusion, the results demonstrate that the prescreening stage can exclude more than 97% non-SDIEs, and the LLM achieves an average accuracy of 95.1% for SDIE classification.

99 - GENERAL AND MISCELLANEOUS↗
DOE OSTI2022

Effectiveness of Deep Learning Trained on SynthCity Data for Urban Point-Cloud Classification

3D object recognition is one of the most popular areas of study in computer vision. Many of the more recent algorithms focus on indoor point clouds, classifying 3D geometric objects, and segmenting outdoor 3D scenes. One of the challenges of the classification pipeline is finding adequate and accurate training data. Hence, this article seeks to evaluate the accuracy of a synthetically generated data set called SynthCity, tested on two mobile laser-scan data sets. Varying levels of noise were applied to the training data to reflect varying levels of noise in different scanners. The chosen deep-learning algorithm was Kernel Point Convolution, a convolutional neural network that uses kernel points in Euclidean space for convolution weights.

Geology↗
DOE OSTI2025

EJFAT Scientific Perspective

Presented new computing model to the test by deploying the EJFAT system alongside a data-stream processing framework running the production-level CLAS12 event reconstruction application. In this experiment, a continuous stream of CLAS12 Level-1 identified events was processed in real-time using the EJFAT load balancer, distributing the workload across 90 computing nodes located across the U.S. This marks the first-ever large-scale, real-time distributed data stream processing experiment, demonstrating that scientific data-streaming pipelines can efficiently scale across four dimensions, thanks to EJFAT’s advanced hardware and software capabilities.

Gyurjyan, Vardan [Thomas Jefferson National Accele↗