Search NASA⌕ Search

SEARCH · Search NASA

Results for “Big data processing”

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 289 records · Page 16
All sourcesNASA NTRSDOE OSTI
NASA NTRSDate not supplied

Towards Automated Analytics of Research Publications

For readers of scientific publications it remains a big challenge to unambiguously relate the published research with the data used. To a substantial degree it is attributed to authors, journals, editors, and reviewers not prioritizing correct data citation, which impacts traceability, repeatability, and giving credits to published authors and their funding sources. Furthermore, uniform classification of the content of the published research is hampered by journals using journal specific topics and letting authors to assign free text keywords to their papers. We demonstrate automated analytics methods for extracting and relating datasets used and the research application areas by processing 1,300 research papers that referenced the NASA Giovanni service (but probably not the datasets in particular) as supporting their publication process. This presentation was given during the 2022 ESIP January meeting held virtually in January 2022.

Irina Gerasimov↗
DOE OSTI2022

Combinatorial Evaluation of Physical Feature Engineering, Classical Machine Learning, and Deep Learning Models for Synchrophasor Data at Scale

A major objective of the project was to train and evaluate the effectiveness of multiple event and anomaly detection, identification and classification deep temporal learning models for processing of real-time phasor measurement unit (PMU) data streams. A vast dataset, consisting of two years of phasor measurements from all three U.S. Interconnections, was curated and released by the Department of Energy (DOE) through Pacific Northwest National Laboratory (PNNL). The dataset also included an event log that provided event times and types (e.g. generator trips, line trips, planned service events, transformer operations, etc.). Our analysis of this dataset addressed six (6) of the eleven (11) research priorities identified in Funding Opportunity Announcement (FOA) DE-FOA-0001861 “Big Data Analysis of Synchrophasor Data” (FOA 1861). Rather than being limited to pre-determined specific algorithms, this project relied on the uniquely structured, highly performant underlying time series database capabilities of the PredictiveGrid platform to assess the vast dataset utilizing a wide variety of algorithms.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2018

FY2018 Discovery Through Situational Awareness: Anomalies, Oscillations, and Classification

This technical report is a summary of work performed under the DOE GMLC program under the task titled “Discovery through Situational Awareness”. The focus of this work was to apply big data, statistical techniques to PMU data to do the following: 1) baseline PMU data and find anomalies with a focus on phase angle pairs, 2) identify the frequencies of dominant modal oscillations between different parts of a power system, and 3) identify and classify frequency events in PMU data. This work was performed on PMU data from the Western Interconnect. Many of these algorithms will be applied to wide area phase angle pair data on the Eastern Interconnect in a tool called ESAMS (Eastern Interconnect Situational Awareness Monitoring System). ESAMS was built by EPG (Electric Power Group) and PNNL and is currently being installed on the Eastern Interconnect. The results discussed in this report were made from 14 PMUs from data ranging from Oct 2016 to May 2018. The PMU data were measured 60 times each second. Fourteen phase angle pairs were included, with reactive and active power being calculated at 14 sites. All results have been de-identified for this report. The methodologies applied in the ESAMS tool are described in Section 2. A baseline of PMU data is created using the last 120 days of data. This baseline describes typical behavior as understood from the data. The baseline is then applied to the next day’s data and anomalies from that baseline are identified. Plots are created to help describe what is anomalous during that specific moment in time. These algorithms are currently being installed on the Eastern Interconnect. FY19 work will include discussing the results and findings from the ESAMS tool. A new method of identifying the frequencies of dominant modal oscillations between different parts of a power system is described in Section 3. A list of 199 frequency events provided by Bonneville Power Administration were used to create algorithms to detect frequency events and to create classification rules that can be used to identify future frequency events. Section 4 discusses the results the performance of 7 different event detection, clustering-based methods. Accuracy percentages ranged from 93.59% to 99.16%, with most methods being very fast. GBM (Gradient Boosting Machine) performed well and quickly. These 7 different methods were also tested to determine how well each built classification rules on training data and then applied those rules to identify future frequency events. Accuracy percentages ranged from 89.09% to 99.9%, with all methods performing quickly. GLM (General Linear Model) was viewed as the top performer when considering accuracy and processing time. Because the list of events consisted of 80% of them as active power events and only 20% as faults, all methods struggled to identify events as faults. Small faults were especially difficult to identify. A separate algorithm was created to help with this. Additional work will be done in FY19 on event detection using more data and more events and refinements to the methods.

24 POWER TRANSMISSION AND DISTRIBUTION↗
NASA NTRS2000

Rovers as Geological Helpers for Planetary Surface Exploration

Rovers can be used to perform field science on other planetary surfaces and in hostile and dangerous environments on Earth. Rovers are mobility systems for carrying instrumentation to investigate targets of interest and can perform geologic exploration on a distant planet (e.g. Mars) autonomously with periodic command from Earth. For nearby sites (such as the Moon or sites on Earth) rovers can be teleoperated with excellent capabilities. In future human exploration, robotic rovers will assist human explorers as scouts, tool and instrument carriers, and a traverse "buddy". Rovers can be wheeled vehicles, like the Mars Pathfinder Sojourner, or can walk on legs, like the Dante vehicle that was deployed into a volcanic caldera on Mt. Spurr, Alaska. Wheeled rovers can generally traverse slopes as high as 35 degrees, can avoid hazards too big to roll over, and can carry a wide range of instrumentation. More challenging terrain and steeper slopes can be negotiated by walkers. Limitations on rover performance result primarily from the bandwidth and frequency with which data are transmitted, and the accuracy with which the rover can navigate to a new position. Based on communication strategies, power availability, and navigation approach planned or demonstrated for Mars missions to date, rovers on Mars will probably traverse only a few meters per day. Collecting samples, especially if it involves accurate instrument placement, will be a slow process. Using live teleoperation (such as operating a rover on the Moon from Earth) rovers have traversed more than 1 km in an 8 hour period while also performing science operations, and can be moved much faster when the goal is simply to make the distance. I will review the results of field experiments with planetary surface rovers, concentrating on their successful and problematic performance aspects. This paper will be accompanied by a working demonstration of a prototype planetary surface rover.

Stoker, Carol↗
DOE OSTI2025

A Data Science and Machine Learning Platform Supporting Large Particle Accelerator Control and Diagnostics Applications Final Report: SBIR Initial Phase II DE-SC0022583

The Machine Learning Data Platform (MLDP) is a product providing full-stack support for data science, Machine Learning, and Artificial Intelligence (ML/AI) applications at particle accelerator and large experimental physics facilities. It supports ML/AI applications from front-end, high-speed acquisition of heterogeneous, time-series data, through data archiving and management, to back-end analysis. The MLDP embodies a “data-science ready” platform for data analysis and ML/AI applications in diagnosis, modelling, control, and optimization of these facilities. It provides data scientists and applications a consistent, datacentric interface to archive data standardizing implementation and deployment of ML/AI algorithms to different operations configurations within the same facility, or between facilities. Being an open-source, public-domain project, the MLDP is intended for broadest possible impact by increasing accessibility and minimizing the required expertise for installation and operation. The MLDP can also be deployed at user facilities for experimental data collection, archiving, and analysis. It is capable of acquisition and archiving of heterogeneous data from experimental equipment (e.g., images, arrays, structures, etc.) along with system hardware configurations (e.g., scalars, tables), control system process variables, and any metadata required for provenance. Thus, the MLDP can manage experimental data through its entire lifecycle, from acquisition and archiving, through analysis and investigation, to release and final publication.

43 PARTICLE ACCELERATORS↗
DOE OSTI2024

Machine Learning-Enabled Image Classification for Automated Electron Microscopy

Abstract Traditionally, materials discovery has been driven more by evidence and intuition than by systematic design. However, the advent of “big data” and an exponential increase in computational power have reshaped the landscape. Today, we use simulations, artificial intelligence (AI), and machine learning (ML) to predict materials characteristics, which dramatically accelerates the discovery of novel materials. For instance, combinatorial megalibraries, where millions of distinct nanoparticles are created on a single chip, have spurred the need for automated characterization tools. This paper presents an ML model specifically developed to perform real-time binary classification of grayscale high-angle annular dark-field images of nanoparticles sourced from these megalibraries. Given the high costs associated with downstream processing errors, a primary requirement for our model was to minimize false positives while maintaining efficacy on unseen images. We elaborate on the computational challenges and our solutions, including managing memory constraints, optimizing training time, and utilizing Neural Architecture Search tools. The final model outperformed our expectations, achieving over 95% precision and a weighted F-score of more than 90% on our test data set. This paper discusses the development, challenges, and successful outcomes of this significant advancement in the application of AI and ML to materials discovery.

Materials Science↗
NASA NTRS2003

Image Processor Electronics (IPE): The High-Performance Computing System for NASA SWIFT Mission

Gamma Ray Bursts (GRBs) are believed to be the most powerful explosions that have occurred in the Universe since the Big Bang and are a mystery to the scientific community. Swift, a NASA mission that includes international participation, was designed and built in preparation for a 2003 launch to help to determine the origin of Gamma Ray Bursts. Locating the position in the sky where a burst originates requires intensive computing, because the duration of a GRB can range between a few milliseconds up to approximately a minute. The instrument data system must constantly accept multiple images representing large regions of the sky that are generated by sixteen gamma ray detectors operating in parallel. It then must process the received images very quickly in order to determine the existence of possible gamma ray bursts and their locations. The high-performance instrument data computing system that accomplishes this is called the Image Processor Electronics (IPE). The IPE was designed, built and tested by NASA Goddard Space Flight Center (GSFC) in order to meet these challenging requirements. The IPE is a small size, low power and high performing computing system for space applications. This paper addresses the system implementation and the system hardware architecture of the IPE. The paper concludes with the IPE system performance that was measured during end-to-end system testing.

Nguyen, Quang H.↗
DOE OSTI2023

Smart Mobility in the Cloud: Enabling Real-Time Situational Awareness and Cyber-Physical Control Through a Digital Twin for Traffic

This article presents the design, implementation, and use cases of the Chattanooga Digital Twin (CTwin) towards the vision for next-generation smart city applications for urban mobility management. CTwin is an end-to-end web-based platform that incorporates various aspects of the decision-making process for optimizing urban transportation systems in Chattanooga, Tennessee, to reduce traffic congestion, incidents, and vehicle fuel consumption. The platform serves as a cyberinfrastructure to collect and integrate multi-domain urban mobility data from various online repositories and Internet of Things (IoT) sensors, covering multiple urban aspects (e.g., traffic, natural hazards, weather, and safety) that are relevant to urban mobility management. The platform enables advanced capabilities for: (a) real-time situational awareness on traffic and infrastructure conditions on highways and urban roads, (b) cyber-physical control for optimizing traffic signal timing, and (c) interactive visual analytics on big urban mobility data and various metrics for traffic prediction and transportation performance evaluation. The platform is designed using a multi-level componentization paradigm and is implemented using modular and adaptive architecture, rendering it as a generalizable and extendable prototype for other urban management applications. We present several use cases to demonstrate CTwin's core capabilities for supporting decision-making in smart urban mobility management.

33 ADVANCED PROPULSION SYSTEMS↗
DOE OSTI2020

Multi-Functional Distributed Fiber Sensors for Pipeline Monitoring and Methane Detections. Final Report

As an abundant and cheap fossil energy source, natural gas has become a significant energy supply to support the United States’ economy. However, the large-scale extraction and utilization of natural gas also impose significant challenges on methane leakage. This problem is exacerbated by aging gas utility delivery systems, including interstate high-pressure pipelines, storage, and transmission facilities. This project aims to develop a cost-effective fiber optical sensing method that can perform multi-parameter real-time measurements of natural gas pipelines across long interrogation distances up to 100 km with 1-meter spatial resolution. This sensing tool can evaluate overall pipeline efficiency and reduce methane emissions for mid-stream methane infrastructures. To accomplish this objective, research and development efforts funded by this project have resulted in the following accomplishments: This project successfully has developed new functional sensory polymer materials using Metal-Organic Frameworks (MOFs) that can be coated on optical fiber through the reel-to-reel coating process. Functional polymer-coated optical fibers can perform sensitive methane detection through evanescence wave interaction and strain-based measurements to achieve 1% detection sensitivities. The new sensors fibers support both distributed measurements and multiplexed fiber sensors array for multi-point measurements. This project developed and optimized a new multi-core optical fiber that supports simultaneous and distributed measurements of strain and temperatures with 1-meter spatial resolutions across up to 100-km interrogation distance. This new fiber, combined with sensory polymercoated fiber, could perform both distributed temperature and methane detections. This project developed a new artificial intelligence big data algorithm approach that can effectively analyze high-resolution data harnessed by distributed fiber sensors to protect natural gas pipelines against external threats and detect internal defects induced by corrosion. Working with our industry partner, this project developed new optical fibers that support fiber sensor fabrications through polymer coating after the fibers are drawn. These new fibers eliminate the need for direct sensor fabrication when the fiber is fabricated on a fiber draw tower, which drastically expands fiber sensors' applicability. This research project has significantly advanced the distributed fiber sensing technology. It will dramatically increase the applicability and adaptability of distributed fiber sensors for a wide array of applications in energy, sustainability, and environmental science, including structural health monitoring of natural gas pipelines, oil infrastructures, hydrogen facilities, and environmental monitoring of carbon storage sites, water supply systems, and others.

03 NATURAL GAS↗
DOE OSTI2020

Adaptive Anomaly Detection for Dynamic Clinical Event Sequences

Over the past decade, health information technology (IT) has enabled the amount of digital information stored in electronic health records (EHRs) to expand greatly. However, according to some studies, hazards in health IT can lead to changes in clinical decisions, care processes, and care outcomes, as well as other issues. Thus, the effects of health IT hazards on patient safety have been at the forefront of recent patient safety research. Nonetheless, hazard detection in health IT remains a challenge. In this paper, the authors assume that safety-related issues in health IT would exhibit anomalous characteristics in EHR data. Although all hazards will exhibit some anomalous characteristics, not all anomalies can be regarded as hazards. The authors hypothesize that errors in health IT could lead to interruptions in the sequence of clinical actions. To this end, the problem of detecting anomalous sequences in big EHR data is considered. This paper focuses on dynamic event sequences, which are a series of clinical actions in motion. The authors propose an adaptive anomaly detection approach that uses higher-order network representation to detect anomalous sequences. Furthermore, the authors propose a contiguous subsequence anomaly detection approach that identifies abnormal subsequences in the detected anomalous sequences. The proposed approaches are tested by using synthetic and real-world EHR data. The proposed methods outperform existing state of the art anomaly detection techniques. To reduce the computational complexity associated with the operational implementation of the proposed approaches, the Apache Spark environment was leveraged, and a much shorter run time together with improved performance were achieved, especially for data with more than 60,000 sequences.

Niu, Haoran↗
DOE OSTI2025

Barge Site - Avian Radar System / Derived Data

This is a combined data set of 67,410 bird/bat tracks from an avian radar system deployed on a research barge (MERLIN True3D, DeTect, Panama City, Florida, USA) and concurrent wind measurements from two scanning lidars (WindCube v2.1, Vaisala, Vantaa, Finland, and Halo XR+, Halo Photonics, Lannion, France). The research barge (16.5 m x 61 m) was deployed as part of the Wind Forecast Improvement Project (WFIP-3) off the northeast coast of the United States south of Massachusetts (40.9 deg N, 70.79 deg W). This data set comprises 5 weeks of data between August 27th 2024 and September 27th 2024. Radar data were provided by DeTect and Lidar data were accessed through the Wind Data Hub (wfip3/barg.WINDPROF.z01.a0) The data have been filtered and sorted into two size groups ("big" and "small") based on a clustering approach. See Snortland, A., Clerc, J., Hein, C., & Cotter, E. (2025). Wind as Driver of Bird and Bat Abundance, Flight Direction, Altitude, and Speed on the North Atlantic Shelf. arXiv preprint arXiv:2511.14983 for complete details. Data are provided in 2 files: "Birds" and "Birds_hourly" Birds: This file contains information about each of the 67,410 flying animal tracks detected by the radar during the data collection period, including parameters measured by the radar and wind information interpolated from the lidar wind measurements. We note that the raw radar dataset contained 301,618 tracks; tracks in this processed dataset were filtered based on the requirements described in Snortland et al. (2025). Birds_hourly: This file contains timeseries of the number of tracks detected per hour over the course of the data collection period, including wind conditions and sun position for each hour. These data were used for generalized additive modeling in Snortland et al. (2025).

17 WIND ENERGY↗
NASA NTRS1988

Massive parallelism in the future of science

Massive parallelism appears in three domains of action of concern to scientists, where it produces collective action that is not possible from any individual agent's behavior. In the domain of data parallelism, computers comprising very large numbers of processing agents, one for each data item in the result will be designed. These agents collectively can solve problems thousands of times faster than current supercomputers. In the domain of distributed parallelism, computations comprising large numbers of resource attached to the world network will be designed. The network will support computations far beyond the power of any one machine. In the domain of people parallelism collaborations among large groups of scientists around the world who participate in projects that endure well past the sojourns of individuals within them will be designed. Computing and telecommunications technology will support the large, long projects that will characterize big science by the turn of the century. Scientists must become masters in these three domains during the coming decade.

Denning, Peter J.↗
DOE OSTI2021

MindSynchro

This report presents the developments and results of MindSynchro project as part of DOE OE FOA 1861. DOE and Pacific Northwest National Laboratory (PNNL) have made available to FOA awardees datasets containing years of real historical data recorded from various phasor measurement units (PMUs) which are installed in three large US interconnections: Texas (IC A), Western (IC B), and Eastern (IC C). The main goal of the project, which was successfully achieved, was to develop methods for detection and identification of events which are relevant for power grid operation. Tasks performed for achieving the project goals included data exploration and pre-processing, the development and application of physics-based features, data analysis and labeling based on unsupervised learning approaches, training and testing of DSSL models for classification of events which are relevant for power grid operation, and deployment of solutions to cloud environments. The methods developed in the project can potentially provide relevant benefits to power grid asset owners/operators in general in terms of situational awareness. Two main types of outcomes can be provided by these tools: Identification of specific relevant power grid event types: Semi-supervised ML methods developed in the project can adequately employ not only the relatively scarce labeled data but also the large amount of available unlabeled data to train models for detection of specific event types. Such methods enable the application of trained models for the detection of events in a population of PMUs much larger than that associated to the labeled events. Support in data labeling / label validation: Labels are critical for training of models for identification of specific types of events. However, labeling large amounts of data is a manual and tedious process. This means that such process is error prone and is not scalable. Methods developed in the project, based on ensembles of clustering models, have been successfully employed for turning manual labeling into a scalable process. Accurate identification of specific relevant events can provide the operators with immediate situational awareness that could otherwise require hours or days of analysis from domain experts. We envision that such methods could be initially employed in support of post-mortem analysis of events and, as confidence is gained, they could be employed for online/real-time support, providing, among other benefits, insights for avoiding major events which could happen due to a combination of smaller ones. On the longer term, related methods could potentially be employed to improve protection and control.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2022

Near real-time streaming analysis of big fusion data

Experiments on fusion plasmas produce high-dimensional data time series with ever-increasing magnitude and velocity, but turn-around times for analysis of this data have not kept up. For example, many data analysis tasks are often performed in a manual, ad-hoc manner some time after an experiment. In this article, we introduce the Delta framework that facilitates near real-time streaming analysis of big and fast fusion data. By streaming measurement data from fusion experiments to a high-performance compute center, Delta allows computationally expensive data analysis tasks to be performed in between plasma pulses. This article describes the modular and expandable software architecture of Delta and presents performance benchmarks of individual components as well as of an example workflow. Focusing on a streaming analysis workflow where electron cyclotron emission imaging (ECEi) data is measured at KSTAR on the National Energy Research Scientific Computing Center's (NERSC's) supercomputer we routinely observe data transfer rates of about 4 Gigabit per second. In NERSC, a demanding turbulence analysis workflow effectively utilizes multiple nodes and graphical processing units and executes them in under 5 min. We further discuss how Delta uses modern database systems and container orchestration services to provide web-based real-time data visualization. For the case of ECEi data we demonstrate how data visualizations can be augmented with outputs from machine learning models. Here, by providing session leaders and physics operators, results of higher-order data analysis using live visualizations may make more informed decisions on how to configure the machine for the next shot.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
NASA NTRS2018

Using Knowledge Analytics to Search and Characterize Mass Properties Aerospace Data

There is growing capability in the field of “Big Data” and “Data Analytics” which Mass Properties Engineers might like to take advantage of. This paper utilizes an implementation of the IBM Knowledge Analytics and Watson search capabilities to explore a corpus of material developed primarily with the interests of Mass Properties Engineers and vehicle concept developers at its forefront. The full collection of SAWE (Society of Allied Weight Engineers, Inc.) Technical Papers from 1939 through 2015 is a major portion of the knowledge content. Additional aerospace vehicle design information includes metadata from AIAA (American Institute for Aeronautics and Astronautics), and INCOSE (International Council on Systems Engineering) as well as author-provided personal search material. This data is processed with respect to certain expected content, data taxonomies and key words to become the core data in NASA Langley Research Center’s “Vehicle Analysis Analytics”, IBM Watson Content. Processed data becomes the corpus of information which is interrogated to provide examples of finding data for mass regression analysis, technology impacts on MPE (Mass Properties Engineering), mass properties control, standards, and other aspects of interest.

Cerro, Jeffrey A.↗
DOE OSTI2019

Astrophysical S -factors, Thermonuclear Rates, and Electron Screening Potential for the 3 He(d,p) 4 He Big Bang Reaction via a Hierarchical Bayesian Model

Here, we developed a hierarchical Bayesian framework to estimate S-factors and thermonuclear rates for the 3 He(d,p) 4 He reaction, which impacts the primordial abundances of 3 He and 7 Li. The available data are evaluated and all direct measurements are taken into account in our analysis for which we can estimate separate uncertainties for systematic and statistical effects. For the nuclear reaction model, we adopt a single-level, two-channel approximation of R-matrix theory, suitably modified to take the effects of electron screening at lower energies into account. In addition to the usual resonance parameters (resonance location and reduced widths for the incoming and outgoing reaction channel), we include the channel radii and boundary condition parameters in the fitting process. Our new analysis of the 3 He(d,p) 4 He S-factor data results in improved estimates for the thermonuclear rates. This work represents the first nuclear rate evaluation using R-matrix theory embedded into a hierarchical Bayesian framework, properly accounting for all known sources of uncertainty. Therefore, it provides a test bed for future studies of more complex reactions.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
NASA NTRS2019

Ask-The-Expert: Minimizing Human Review for Big Data Analytics Through Active Learning

In this CIF project, we worked toward semi-automating knowledge discovery from anomaly detection algorithms through the use of active learning. Active learning is an area of research within machine learning that uses an "expert in the loop" to learn from large data sets that have very few annotations or labels available, and where providing such labels is expensive. In our case, the task can be defined as the identification of safety events from flight operational data. Since traditional anomaly detection algorithms cannot differentiate between operationally relevant and irrelevant statistical anomalies, Subject Matter Experts (SMEs) have a lengthy and expensive burden of investigating every example identified by the detection algorithm, classifying and labeling them as relevant or irrelevant. Active learningidentifies the unlabeled example for which a label would most improve the classifier, asks the domain expert for a label, and repeats this process until there are no more resources (time, budget) available for labeling or a minimum required performance is reached. A positive label indicates an operationally significant safety event whereas a negative label indicates otherwise. Based on these few labels we propose to build an active learning system that utilizes the SME's time in the most effective manner by iteratively asking for labels for as few informative instances as possible. Our work was proposed to be a stepping stone toward implementation and deployment of the system with user interface to be pursued by the Aviation Operations and Safety Program (AOSP) given its interest in safety monitoring and discovery of safety incidents.

aviation safety↗
DOE OSTI2020

DeepFreeze: Towards Scalable Asynchronous Checkpointing of Deep Learning Models

In the age of big data, deep learning has emerged as a powerful tool to extract insight and exploit its value, both in industry and scientific applications. One common pattern emerging in such applications is frequent checkpointing of the state of the learning model during training, needed in a variety of scenarios: analysis of intermediate states to explain features and correlations with training data, exploration strategies involving alternative models that share a common ancestor, knowledge transfer, resilience, etc. However, with increasing size of the learning models and popularity of distributed data-parallel training approaches, simple checkpointing techniques used so far face several limitations: low serialization performance, blocking I/O, stragglers due to the fact that only a single process is involved in checkpointing. This paper proposes a checkpointing technique specifically designed to address the aforementioned limitations, introducing efficient asynchronous techniques to hide the overhead of serialization and I/O, and distribute the load over all participating processes. Experiments with two deep learning applications (CANDLE and ResNet) on a pre-Exascale HPC platform (Theta) shows significant improvement over state-of-art, both in terms of checkpointing duration and runtime overhead.

checkpointing↗