Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “small dataset”

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 109 records · Page 6
All sourcesNASA NTRSDOE OSTI
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 OSTI2024

Neural network emulation of flow in heavy-ion collisions at intermediate energies

Applications of new techniques in machine learning are speeding up progress in research in various fields. In this work, we construct and evaluate a deep neural network (DNN) to be used within a Bayesian statistical framework as a faster and more reliable alternative to the Gaussian process (GP) emulator of an isospin-dependent Boltzmann-Uehling-Uhlenbeck (IBUU) transport model simulator of heavy-ion reactions at intermediate beam energies. We found strong evidence of the DNN being able to emulate the IBUU simulator's prediction on the strengths of protons' directed and elliptical flow very efficiently even with small training datasets and with accuracy about ten times higher than the GP. Here, limitations of our present work and future improvements are also discussed.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2025

Leveraging Pre-Built Catalogs and Object-Level Scheduling to Eliminate I/O Bottlenecks in HPC Environments

Modern High-Performance Computing (HPC) environments face mounting challenges due to the shift from large to small file datasets, along with an increasing number of users and parallelized applications. As HPC systems rely on Parallel File Systems (PFS), such as Lustre for data processing, performance bottlenecks stemming from Object Storage Target (OST) contention have become a significant concern. Existing solutions, such as LADS with its object-level scheduling approach, fall short in large-scale HPC environments due to their inability to effectively address metadata I/O bottlenecks and the growing number of I/O processes. This study highlights the pressing need for a comprehensive solution that tackles both OST contention and metadata I/O challenges in diverse HPC workloads. To address these challenges, we propose SwiftLoad, an object-level I/O scheduling framework that leverages a metadata catalog to enhance the performance and efficiency of parallel HPC utilities. The adoption of the metadata catalog mitigates the metadata I/O bottlenecks that commonly occur in HPC utilities, a challenge that is particularly pronounced in object-level I/O scheduling. SwiftLoad addresses OST contention and the uneven distribution of I/O processes across different OSTs through mathematical modeling and incorporates a Loader Configuration Module to regulate the number of I/O processes. Evaluated with two representative utilities—data deduplication profiling and data augmentation—SwiftLoad achieved performance improvements of up to 5.63x and 11.0x, respectively, on a production supercomputer.

HPC↗
DOE OSTI2022

Descriptor Aided Bayesian Optimization for Many-Level Qualitative Variables With Materials Design Applications

Abstract Engineering design often involves qualitative and quantitative design variables, which requires systematic methods for the exploration of these mixed-variable design spaces. Expensive simulation techniques, such as those required to evaluate optimization objectives in materials design applications, constitute the main portion of the cost of the design process and underline the need for efficient search strategies—Bayesian optimization (BO) being one of the most widely adopted. Although recent developments in mixed-variable Bayesian optimization have shown promise, the effects of dimensionality of qualitative variables have not been well studied. High-dimensional qualitative variables, i.e., with many levels, impose a large design cost as they typically require a larger dataset to quantify the effect of each level on the optimization objective. We address this challenge by leveraging domain knowledge about underlying physical descriptors, which embody the physics of the underlying physical phenomena, to infer the effect of unobserved levels that have not been sampled yet. We show that physical descriptors can be intuitively embedded into the latent variable Gaussian process approach—a mixed-variable GP modeling technique—and used to selectively explore levels of qualitative variables in the Bayesian optimization framework. This physics-informed approach is particularly useful when one or more qualitative variables are high dimensional (many-level) and the modeling dataset is small, containing observations for only a subset of levels. Through a combination of mathematical test functions and materials design applications, our method is shown to be robust to certain types of incomplete domain knowledge and significantly reduces the design cost for problems with high-dimensional qualitative variables.

Engineering↗
DOE OSTI2019

Towards Class-Balancing Human Comfort Datasets with GANs

Human comfort datasets are widely used in smart buildings. From thermal comfort prediction to personalized indoor environments, labelled subjective responses from participants in an experiment are required to feed different machine learning models. However, many of these datasets are small in samples per participants, number of participants, or suffer from a class-imbalance of its subjective responses. In this paper we explore the use of Generative Adversarial Networks to generate synthetic samples to be used in combination with real ones for data-driven applications in the built environment.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2026

ChatMPI: LLM-Driven MPI Code Generation for HPC Workloads

The Message Passing Interface (MPI) standard plays a crucial role in enabling scientific applications for parallel computing and is an essential component in high-performance computing (HPC). However, implementing MPI code manually—especially applying a proper domain decomposition and communication pattern—is a challenging and error-prone task. We present ChatMPI, an AI assistant for MPI parallelization of sequential C codes. In our analysis, we focus on testing six essential HPC workloads, which are based on Basic Linear Algebra Subprograms levels 1, 2, and 3 as well as sparse, stencil, and iterative operations. We analyze the process of creating ChatMPI by using the ChatHPC library. This lightweight large language model (LLM)–based infrastructure enables HPC experts to efficiently create and supervise trustworthy AI capabilities for critical HPC software tasks. We study the data required for training (fine-tuning) ChatMPI to generate parallel codes that not only use MPI syntax correctly but also apply HPC techniques to reduce memory communication and maximize performance by using proper work decomposition. With a relatively small training dataset composed of a few dozen prompts and fewer than 15 minutes of fine-tuning on one node equipped with two NVIDIA H100 GPUs, ChatMPI elevates trustworthiness for MPI code generation of current LLMs (e.g., Code Llama, ChatGPT-4o and ChatGPT 5). Additionally, we evaluate the performance of the MPI codes generated by ChatMPI in comparison with the ones generated by ChatGPT-4o and ChatGPT-5. The codes generated by ChatMPI provide up to a 4 × boost in performance by using better problem decomposition, communication patterns, and HPC techniques (e.g., communication avoiding).

Valero Lara, Pedro [ORNL] (ORCID:0000000214794310)↗
DOE OSTI2026

RxnRover/amlro

AMLRO (Active Machine Learning Reaction Optimizer) is an open-source framework designed to accelerate chemical reaction optimization using active learning with classical machine learning regression models. AMLRO integrates space-filling sampling strategies (e.g., Sobol and Latin Hypercube sampling) with iterative model training, prediction, and experiment selection to efficiently navigate complex reaction spaces. The platform supports multiple regression models, flexible multi-objective definitions, and user-defined parameter bounds, enabling data-efficient optimization from small initial datasets. AMLRO is designed for ease of use by experimentalists and can operate as a standalone decision-support tool or be integrated into closed-loop automated experimentation workflows.

Kulathunga, Dulitha Prasanna [Iowa State Universit↗
NASA NTRS2008

Material Density Distribution of Small Debris in Earth Orbit

Over 200 spacecraft and rocket body breakups in Earth orbit have populated that regime with debris fragments in the sub-micron through meter size range. Though the largest debris fragments can cause significant collisional damage to active (operational) spacecraft, these are few and trackable by radar. Fragments on the order of a millimeter to a centimeter in size are as yet untrackable. But this smaller debris can result in damage to critical spacecraft systems and, under the worst conditions, fragmenting collision events. Ongoing research at the NASA Orbital Debris Program Office on the sources of these small fragments has focused on the material components of spacecraft and rocket bodies and on breakup event morphology. This has led to fragment material density estimates, and also the beginnings of shape categorizations. To date the NASA Standard Breakup Model has not considered specific material density distinctions of small debris. The basis of small debris in that model is the fourth hypervelocity impact event of the Satellite Orbital Debris Characterization Impact Test (SOCIT) series. This test targeted a flight-ready, U.S. Transit navigation satellite with a solid aluminum sphere impactor. Results in this event yield characteristic length (size) and area-to-mass distributions of fragments smaller than 10 cm in the NASA model. Recent re-analysis of the SOCIT4 small fragment dataset highlighted the material-specific characteristics of metals and non-metals. Concurrent analysis of Space Shuttle in-situ impact data showed a high percentage of aluminum debris in shuttle orbit regions. Both analyses led to the definition of three main on-orbit debris material density categories -low density (< 2 g/cc), medium density (2 to 6 g/cc), and high density (> 6 g/cc). This report considers the above studies in an explicit extension of the NASA Standard Breakup Model where separate material densities for debris are generated and these debris fragments are propagated in Earth orbit. The near Earth environment is thus parameterized by debris density percentages within subsections of that environment. This model version is used in the upgraded NASA Orbital Debris Engineering Model (ORDEM).

Krisko, P. H.↗
NASA NTRS2010

A Simple Technique for Creating Regional Composites of Sea Surface Temperature from MODIS for Use in Operational Mesoscale NWP

This paper describes a simple technique for creating regional, high-resolution, daytime and nighttime composites of sea surface temperature (SST) for use in operational numerical weather prediction (NWP). The composites are based on observations from NASA s Moderate Resolution Imaging Spectroradiometer (MODIS) aboard Aqua and Terra. The data used typically are available nearly in real time, are applicable anywhere on the globe, and are capable of roughly representing the diurnal cycle in SST. The composites resolution is much higher than that of many other standard SST products used for operational NWP, including the low- and high-resolution Real-Time Global (RTG) analyses. The difference in resolution is key because several studies have shown that highly resolved SSTs are important for driving the air sea interactions that shape patterns of static stability, vertical and horizontal wind shear, and divergence in the planetary boundary layer. The MODIS-based composites are compared to in situ observations from buoys and other platforms operated by the National Data Buoy Center (NDBC) off the coasts of New England, the mid-Atlantic, and Florida. Mean differences, mean absolute differences, and root-mean-square differences between the composites and the NDBC observations are all within tenths of a degree of those calculated between RTG analyses and the NDBC observations. This is true whether or not one accounts for the mean offset between the skin temperatures of the MODIS dataset and the bulk temperatures of the NDBC observations and RTG analyses. Near the coast, the MODIS-based composites tend to agree more with NDBC observations than do the RTG analyses. The opposite is true away from the coast. All of these differences in point-wise comparisons among the SST datasets are small compared to the 61.08C accuracy of the NDBC SST sensors. Because skin-temperature variations from land to water so strongly affect the development and life cycle of the sea breeze, this phenomenon was chosen for demonstrating the use of the MODIS-based composite in an NWP model. A simulated sea breeze in the vicinity of New York City and Long Island shows a small, net, but far from universal improvement when MODIS-based composites are used in place of RTG analyses. The timing of the sea breeze s arrival is more accurate at some stations, and the near-surface temperature, wind, and humidity within the breeze are more realistic.

Knievel, Jason C.↗
NASA NTRS2020

Object and Gas Source Detection with Robotic Platforms in Perceptually-Degraded Environments

In exploration-oriented robotic missions for disaster relief in unknown subterranean environments, it is of prime importance for a human supervisor to rapidly gain situational awareness of salient objects within the environment. In this paper we present an automated object detection pipeline that is adaptable to heterogeneous robots with arbitrary sensor configurations. It has been deployed in time-critical scenarios with multiple collaborative robots in a variety of demanding underground environments. For visually observable objects, detections are made in both the visible and thermal spectra using a state-of-the-art machine learning framework for object detection and classification. Our pipeline can be rapidly adapted to a specific task by using a small, structured dataset to fine-tune a pre-trained convolutional neural network (CNN). Relative localization is separated from the CNN for speed of operation. A robust architecture for localization is used with outlier rejection and a hierarchy of fall-back distance measurement methods. Point-source objects such as gas and WiFi hotspots can also be detected, by tracking signal strength over time and presenting an intuitive visualization on a map. Observations of each object types are presented to the operator in ranked confidence order for final evaluation.

Agha-mohammadi, Ali-akbar↗
NASA NTRSDate not supplied

Transcriptomics-based Machine Learning (ML) Analysis Predicts Space-Exposed Murine Livers

NASA has employed high-throughput molecular assays to identify sub-cellular changes impacting human physiology during spaceflight. Machine learning (ML) methods hold the promise to improve our ability to identify important signals within highly dimensional molecular data. However, the inherent limitation of study subject numbers within a spaceflight mission minimizes the utility of ML approaches. To overcome the sample power limitations, data from multiple spaceflight missions must be aggregated while appropriately addressing intra- and inter-study variabilities. Here we describe an approach to log transform, scale and normalize data from six heterogeneous, mouse liver derived transcriptomics datasets (ntotal=137) which enabled ML-methods to perform well (AUC ≥ 0.87) in classifying spaceflown vs ground control animals rather than mission-of-origin. Concordance was found between liver-specific biological processes identified from harmonized ML-based analysis and study-by-study classical omics analysis. This work demonstrates the feasibility of applying ML methods on integrated, heterogeneous datasets of small sample size.

Machine Learning↗
NASA NTRSDate not supplied

Document Classification Techniques for Aviation Letters of Agreement

Often when working with technical documents, it is helpful to classify them into specific categories. In this paper, we conduct a thorough review of natural language processing techniques to perform this classification task on Letters of Agreement (LOAs), technical aviation documents outlining rules for utilizing US airspace. We evaluate multiple techniques, including Transfer Learning, for representing the text in the documents as embeddings: unigram and bigram Term Frequency Inverse Document Frequency (TFIDF), Word2Vec, Doc2Vec, GloVe and RoBERTa. We investigate a wide range of classification models: K-Nearest Neighbors, Random Forest, Support Vector Machines (SVM), Logistic Regression, Naive Bayes, Feed-Forward Neural Network, Convolutional Neural Networks (CNNs) and Long-Short Term Memory (LSTM). By comparing the different methods, we found the best overall approach for our task was to use unigram TFIDF representations with SVM while also gaining insight into how the other methodologies performed on a small technical datasets.

Aayushi Batra↗
NASA NTRSDate not supplied

Document Classification Techniques for Aviation Letters of Agreement

Often when working with historic air traffic management (ATM) documents, it is helpful to classify them into specific categories. In this paper, we conduct a thorough review of natural language processing techniques to perform this classification task on Letters of Agreement (LOAs), technical aviation documents outlining rules for utilizing US airspace. We evaluate multiple techniques for representing the text in the documents as embeddings: unigram and bigram Term Frequency Inverse Document Frequency (TFIDF), Word2Vec, Doc2Vec, GloVe and RoBERTa. We investigate a wide range of classification models: K-Nearest Neighbors, Random Forest, Support Vector Machines (SVM), Logistic Regression, Naive Bayes, Feed-Forward Neural Network, Convolutional Neural Networks (CNNs) and Long-Short Term Memory (LSTM). By comparing the different methods, we found the best overall approach for our task was to use unigram TFIDF representations with SVM while also gaining insight into how the other methodologies performed on a small technical datasets.

ATM↗
NASA NTRSDate not supplied

Document Classification Techniques for Aviation Letters of Agreement

Often when working with historic air traffic management (ATM) documents, it is helpful to classify them into specific categories. In this paper, we conduct a thorough review of natural language processing techniques to perform this classification task on Letters of Agreement (LOAs), technical aviation documents outlining rules for utilizing US airspace. We evaluate multiple techniques for representing the text in the documents as embeddings: unigram and bigram Term Frequency Inverse Document Frequency (TFIDF), Word2Vec, Doc2Vec, GloVe and RoBERTa. We investigate a wide range of classification models: K-Nearest Neighbors, Random Forest, Support Vector Machines (SVM), Logistic Regression, Naive Bayes, Feed-Forward Neural Network, Convolutional Neural Networks (CNNs) and Long-Short Term Memory (LSTM). By comparing the different methods, we found the best overall approach for our task was to use unigram TFIDF representations with SVM while also gaining insight into how the other methodologies performed on a small technical datasets.

ATM↗
DOE OSTI2014

FASTERp: A Feature Array Search Tool for Estimating Resemblance of Protein Sequences

Metagenome sequencing efforts have provided a large pool of billions of genes for identifying enzymes with desirable biochemical traits. However, homology search with billions of genes in a rapidly growing database has become increasingly computationally impractical. Here we present our pilot efforts to develop a novel alignment-free algorithm for homology search. Specifically, we represent individual proteins as feature vectors that denote the presence or absence of short kmers in the protein sequence. Similarity between feature vectors is then computed using the Tanimoto score, a distance metric that can be rapidly computed on bit string representations of feature vectors. Preliminary results indicate good correlation with optimal alignment algorithms (Spearman r of 0.87, ~;;1,000,000 proteins from Pfam), as well as with heuristic algorithms such as BLAST (Spearman r of 0.86, ~;;1,000,000 proteins). Furthermore, a prototype of FASTERp implemented in Python runs approximately four times faster than BLAST on a small scale dataset (~;;1000 proteins). We are optimizing and scaling to improve FASTERp to enable rapid homology searches against billion-protein databases, thereby enabling more comprehensive gene annotation efforts.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2020

A Scalable Parallel Hypergraph Generator (HyGen)

Graphs are extensively used to model real-world complex systems. An edge in a graph can model pairwise relationships. However, multiway relationships (connections between three or more vertices) are common in many complex systems such as cellular process, image segmentation, and circuit design. A graph edge cannot model multiway relationships. A hypergraph, which can connect more than two vertices, is thus a better option to model multiway relationships. A large-scale hypergraph analysis has the potential to find useful insights from a complex system and assist in knowledge discovery. Currently a limited number of hypergraphs exists that are representative of real-world datasets. Moreover, real-world hypergraph datasets are small in size and inadequate to incorporate future needs. A graph generator that can produce large-scale synthetic hypergraphs can solve the above mentioned problems. In this paper, we present a scalable parallel hypergraph generator (HyGen) based on the Message Passing Interface (MPI) standard. To generate hypergraphs, HyGen takes the following parameter values as inputs: i) number of vertices, ii) number of hyperedges, iii) number of clusters, iv) vertex distribution, v) hyperedge distribution, vi) local cluster cardinality, and vii) global cluster cardinality. We have demonstrated that HyGen can generate hypergraphs of various sizes in a scalable fashion. HyGen takes approximately four minutes to generate a hypergraph with 4.8 million vertices, 1.6 million hyperedges, and 800 clusters using 1,024 processes on a leadership class computing platform. Our strong and weak scaling experiments on supercomputers demonstrate that HyGen can quickly create large-scale hypergraphs in a parallel manner, thus providing a useful capability for hypergraph analysis.

Hasan, S M Shamimul↗
DOE OSTI2019

Cs-137 Body-burdens in Wild Pigs on the Department of Energy's Savannah River Site: Determination of an Appropriate Framework for Spatiotemporal Analysis of Hunter Dose Tracking System Data

The Savannah River Site (SRS) is a U.S. Department of Energy facility with an area of 310 square miles. Since its opening in 1951, SRS has produced and processed nuclear materials for national defense and industrial purposes. These activities have resulted in the release of radiocesium (Cs-137) which has led to contamination of associated streams, floodplains, man-made reservoirs, soil and vegetation. Contamination is spread to wild pigs as they move about SRS, consume vegetation, drink water and wallow in soil. SRS provides annual deer hunts open to the public in which wild pigs may be taken as well. The Cs-137 concentration of each animal is determined before it is released to the hunter for consumption and a cumulative dose is retained for each individual hunter. These tasks are accomplished using the Hunter Dose Tracking System (HDTS). The data collected on SRS by the HDTS for White-Tailed deer has been extensively analyzed. A repeatable method for calculating the background radiocesium body burden for the SRS deer population was established by Gaines and Novak (2009,2016). This model was framed using Cs-137 contamination obtained via gamma over-flight data, the decay rate of radiocesium and the historical behavior observed in the SRS deer population. These parameters were used in a spatiotemporal analysis to estimate background Cs-137 body burdens (from global fallout, etc.), as well as the effect SRS activities may have. The purpose of this project is to modify the framework to represent the behavior of the wild pig population on SRS such that HDTS data collected may be analyzed similarly. The 2018 data were collected via the HDTS in the field to screen wild pigs dispatched by Forest Service - Savannah River. These data will provide insight for an analysis of seasonal variance in the body burdens of the SRS wild pig population. The HDTS quickly provided Cs-137 concentrations in the field, as well as automatically updated the electronic database. Space: Historically, the average home range of wild pigs on the Savannah River Site has been reported at approximately six square miles. This range was confirmed accurate as of July 2018, with the observed minimum and maximum at one and ten square miles respectively. This range was geoprocessed using ArcMap to project buffers based on 1991 gamma over-flight data. The buffers were constructed to originate at all Cs-137 contamination point coordinates and protrude outward in all directions for 1.5, 2.45 and 6 square miles. While the 2.45 mile buffer is theoretically representative of potential hog exposure based on the average home range, it does not consider the possibility of home ranges not uniform in shape. Therefore, the 1.5 and 6 mile buffers were also constructed as visual aids for determining the appropriate hunt compartments for background level considerations. The SRS is divided into numbered hunt compartments of varying size and shape. These compartment numbers are used during sponsored hunts to record the location in which a wild pig was dispatched. Time: The Gaines report determined deer analysis should be carried out at the half-decade level for the best representation of true body burdens. This is due to the impact natural factors, such as drought or flood, can have on the bioaccumulation of Cs-137. To determine the appropriate time scale for hog analysis, all available data was totaled and averaged at 1, 2, 3 and 7 year increments. The annual averages and running average are provided. The appropriate compartments for the determination of the SRS hog population background level should be: 2, 3, 4, 9, 10, 14 and 15. Hogs dispatched within other compartments are most likely exposed to Cs-137 contamination and should not be included in the background level. The current dataset is small, but was still able to demonstrate the relationship between where hogs on SRS and the Cs-137 body burdens they incur. Geoprocessing the data using the average home range of wild pigs highlighted the importance of collecting additional samples from areas surrounding SRS if an accurate background is to be determined. When the average concentration of Cs-137 in pigs from background compartments is compared against that of pigs from contaminated compartments, a significant difference can be seen. A Table provides a summary of the analyzed data from the 15 years processed thus far.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗