Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Distributed training”

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 631 records · Page 35
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

I/O in Machine Learning Applications on HPC Systems: A 360-degree Survey

Growing interest in Artificial Intelligence (AI) has resulted in a surge in demand for faster methods of Machine Learning (ML) model training and inference. This demand for speed has prompted the use of high performance computing (HPC) systems that excel in managing distributed workloads. Because data is the main fuel for AI applications, the performance of the storage and I/O subsystem of HPC systems is critical. In the past, HPC applications accessed large portions of data written by simulations or experiments or ingested data for visualizations or analysis tasks. ML workloads perform small reads spread across a large number of random files. This shift of I/O access patterns poses several challenges to modern parallel storage systems. In this paper, we survey I/O in ML applications on HPC systems, and target literature within a 6-year time window from 2019 to 2024. We define the scope of the survey, provide an overview of the common phases of ML, review available profilers and benchmarks, examine the I/O patterns encountered during offline data preparation, training, and inference, and explore I/O optimizations utilized in modern ML frameworks and proposed in recent literature. Lastly, we seek to expose research gaps that could spawn further R&D.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS1991

Blood Volume: Its Adaptation to Endurance Training

Expansion of blood volume (hypervolemia) has been well documented in both cross-sectional and longitudinal studies as a consequence of endurance exercise training. Plasma volume expansion can account for nearly all of the exercise-induced hypervolemia up to 2-4 wk; after this time expansion may be distributed equally between plasma and red cell volumes. The exercise stimulus for hypervolemia has both thermal and nonthermal components that increase total circulating plasma levels of electrolytes and proteins. Although protein and fluid shifts from the extravascular to intravascular space may provide a mechanism for rapid hypervolemia immediately after exercise, evidence supports the notion that chronic hypervolemia associated with exercise training represents a net expansion of total body water and solutes. This net increase of body fluids with exercise training is associated with increased water intake and decreased urine volume output. The mechanism of reduced urine output appears to be increased renal tubular reabsorption of sodium through a more sensitive aldosterone action in man. Exercise training-induced hypervolemia appears to be universal among most animal species, although the mechanisms may be quite different. The hypervolemia may provide advantages of greater body fluid for heat dissipation and thermoregulatory stability as well as larger vascular volume and filling pressure for greater cardiac stroke volume and lower heart rates during exercise.

Convertino, Victor A.↗
DOE OSTI2024

Unsupervised Learning for Equitable DER Control: Preprint

In the context of managing distributed energy resources (DERs) within distribution networks (DNs), this work focuses on the task of developing local controllers. We propose an unsupervised learning framework to train functions that can closely approximate optimal power flow (OPF) solutions. The primary aim is to establish specific conditions under which these learned functions can collectively guide the network towards desired configurations asymptotically, leveraging an incremental control approach. The flexibility of the proposed methodology allows to integrate fairness-driven components into the cost function associated with the OPF problem. This addition seeks to mitigate power curtailment disparities among DERs, thereby promoting equitable power injections across the network. To demonstrate the effectiveness of the proposed approach, power flow simulations are conducted using the IEEE 37-bus feeder. The findings not only showcase the guaranteed system stability but also underscore its improved overall performance.

asymptotic stability↗
DOE OSTI2024

Equipping Neural Network Surrogates with Uncertainty for Propagation in Physical Systems

Coarse-grained or filtered models typically rely on closure models to account for unresolved scales. For instance, large eddy simulation for modeling turbulent fluid flows explicitly resolves the largest scales, but requires modeling closure terms to account for the sub-filter scales. With the vast amount of data available from high-fidelity simulations, there are unique opportunities to leverage data-driven modeling techniques to formulate expressive and flexible closure models. Despite their flexibility, data-driven models struggle in domain shift settings, i.e. when deployed in configurations not captured in the training dataset. In particular, the efficacy of neural network surrogates is difficult to assess a priori due to the deterministic, point-estimate nature of predictions. In high-consequence applications, such models require reliable uncertainty estimates in the data-informed and out-of-distribution regimes. To quantify uncertainties in both regimes, we employ Bayesian neural networks which are able to capture both epistemic and aleatoric uncertainties. We will discuss challenges associated with the training and evaluation of these networks. Furthermore, we will discuss uncertainty embedding strategies to enable efficient sampling and propagation of uncertainty through high-fidelity simulations.

Bayesian neural networks↗
DOE OSTI2021

A machine learning approach to galaxy properties: joint redshift–stellar mass probability distributions with Random Forest

We demonstrate that highly accurate joint redshift–stellar mass probability distribution functions (PDFs) can be obtained using the Random Forest (RF) machine learning (ML) algorithm, even with few photometric bands available. As an example, we use the Dark Energy Survey (DES), combined with the COSMOS2015 catalogue for redshifts and stellar masses. We build two ML models: one containing deep photometry in the griz bands, and the second reflecting the photometric scatter present in the main DES survey, with carefully constructed representative training data in each case. We validate our joint PDFs for 10 699 test galaxies by utilizing the copula probability integral transform and the Kendall distribution function, and their univariate counterparts to validate the marginals. Benchmarked against a basic set-up of the template-fitting code bagpipes, our ML-based method outperforms template fitting on all of our predefined performance metrics. In addition to accuracy, the RF is extremely fast, able to compute joint PDFs for a million galaxies in just under 6 min with consumer computer hardware. Such speed enables PDFs to be derived in real time within analysis codes, solving potential storage issues. As part of this work we have developed galpro 1, a highly intuitive and efficient python package to rapidly generate multivariate PDFs on-the-fly. galpro is documented and available for researchers to use in their cosmology and galaxy evolution studies.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2021

ThickBrick: optimal event selection and categorization in high energy physics. Part I. Signal discovery

We provide a prescription called ThickBrick to train optimal machine-learning-based event selectors and categorizers that maximize the statistical significance of a potential signal excess in high energy physics (HEP) experiments, as quantified by any of six different performance measures. For analyses where the signal search is performed in the distribution of some event variables, our prescription ensures that only the information complementary to those event variables is used in event selection and categorization. This eliminates a major misalignment with the physics goals of the analysis (maximizing the significance of an excess) that exists in the training of typical ML-based event selectors and categorizers. In addition, this decorrelation of event selectors from the relevant event variables prevents the background distribution from becoming peaked in the signal region as a result of event selection, thereby ameliorating the challenges imposed on signal searches by systematic uncertainties. Our event selectors (categorizers) use the output of machine-learning-based classifiers as input and apply optimal selection cutoffs (categorization thresholds) that are functions of the event variables being analyzed, as opposed to flat cutoffs (thresholds). These optimal cutoffs and thresholds are learned iteratively, using a novel approach with connections to Lloyd’s k-means clustering algorithm. We provide a public, Python implementation of our prescription, also called ThickBrick, along with usage examples.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2020

Ultrasonic Measurements of Temperature Distribution in Extreme Environments: Results of Power Plant Testing

The most hardened insertion sensors do not perform reliably in extreme environmental conditions. We address this problem by developing an ultrasonic thermometry system for the measurements of segmental temperature distributions (MSTD) in solid materials and, specifically, across containments of energy conversion processes. The MSTD approach uses an echogenically segmented ultrasound propagation path, structured to contain engineered or naturally occurring echogenic features. In pulse-echo mode, these features produce a train of echoes in response to an external pulse of ultrasonic excitation created and acquired by a transducer located outside the extreme environment. The delays between the echoes encode the information on the temperature distribution in the corresponding segments of the ultrasonic propagation path, which we use to reconstruct an unknown temperature profile. This note describes the preliminary results and initial experience in applying the developed MSTD ultrasonic thermometry system to the measurements of temperatures in a utility-scale 500MW coal-fired boiler.

John, Mason↗
DOE OSTI2023

DGaaS: GPU as a Service on Distributed Computing System

In the rapidly evolving landscape of scientific computing, Graphics Processing Units (GPUs) have become indispensable for their unparalleled ability to handle parallel tasks in complex calculations, simulations, and data analysis. Their utility is further magnified in machine learning and AI applications, where they significantly accelerate model training and predictive analytics. Within this context, the Triton Inference Server emerges as a pivotal open-source tool, specializing in AI inferencing and optimizing GPU utilization across various platforms and frameworks. This paper presents an in-depth study on distributed High Throughput Computing (HTC), specifically focusing on the HTCondor framework and its resource provisioning tools, GlideinWMS and HEPCloud. These systems enable large-scale scientific experiments like CMS and DUNE to efficiently access and utilize vast computational resources. The paper explores the core architectural components of GlideinWMS, including jobs, user pools, and worker nodes, and discusses their integration with GPUs and the Triton server. The primary aim of this research is to develop a solution that optimizes GPU utilization by leveraging Glideins and containers. This approach allows computational jobs, particularly those involving AI models, to use GPUs only when essential, thereby facilitating efficient sharing of limited GPU resources. To validate this architecture, the study conducted three key tests involving custom scripts, container-based servers, and Triton server deployments. However, the study faces challenges, notably in locating the Triton server and ensuring secure remote access. To address these issues, future work will focus on developing a proxy mechanism and enhancing security protocols. In conclusion, this study offers a comprehensive roadmap for effective and efficient GPU utilization in distributed High Throughput Computing. It aims to contribute significantly to the scientific community by solving pressing problems and implementing robust solutions in collaboration with the GlideinWMS and HEPCloud teams. The research sets the stage for a more efficient, scalable, and cost-effective paradigm in scientific computing.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Primal-Dual Differentiable Programming for Distribution System Critical Load Restoration: Preprint

Swift and reliable critical load restoration (CLR) can help make a distribution system resilient towards extreme events. To optimally achieve that, alongside practical concerns such as limiting online computational burden, some studies leverage model-free reinforcement learning (RL) to train control policies. Despite the advantages provided by RL algorithms, these approaches suffer from two issues: 1) the lack of a proper mechanism for constraint enforcement, and 2) poor sample efficiency. Therefore, in this paper, a primal-dual differentiable programming (PDDP) method is developed for guiding the training leading to a constraint-satisfying policy. Additionally, the model-based nature of the proposed method aims at improving sample efficiency. The experiment on a CLR problem demonstrates that PDDP can effectively train a control policy that both achieves desirable performance and satisfies required constraints.

differentiable programming↗
DOE OSTI2022

A Hybrid Reinforcement Learning-MPC Approach for Distribution System Critical Load Restoration

This paper proposes a hybrid control approach for distribution system critical load restoration, combining deep reinforcement learning (RL) and model predictive control (MPC) aiming at maximizing total restored load following an extreme event. RL determines a policy for quantifying operating reserve requirements, thereby hedging against uncertainty, while MPC models grid operations incorporating RL policy actions (i.e., reserve requirements), renewable (wind and solar) power predictions, and load demand forecasts. We formulate the reserve requirement determination problem as a sequential decision-making problem based on the Markov Decision Process (MDP) and design an RL learning environment based on the OpenAI Gym framework and MPC simulation. The RL agent reward and MPC objective function aim to maximize and monotonically increase total restored load and minimize load shedding and renewable power curtailment. The RL algorithm is trained offline using a historical forecast of renewable generation and load demand. The method is tested using a modified IEEE 13-bus distribution test feeder containing wind turbine, photovoltaic, microturbine, and battery. Case studies demonstrated that the proposed method outperforms other policies with static operating reserves.

distribution system↗
DOE OSTI2025

Systematic softening in universal machine learning interatomic potentials

Machine learning interatomic potentials (MLIPs) have introduced a new paradigm for atomic simulations. Recent advancements have led to universal MLIPs (uMLIPs) that are pre-trained on diverse datasets, providing opportunities for universal force fields and foundational machine learning models. However, their performance in extrapolating to out-of-distribution complex atomic environments remains unclear. In this study, we highlight a consistent potential energy surface (PES) softening effect in three uMLIPs: M3GNet, CHGNet, and MACE-MP-0, which is characterized by energy and force underprediction in atomic-modeling benchmarks including surfaces, defects, solid-solution energetics, ion migration barriers, phonon vibration modes, and general high-energy states. The PES softening behavior originates primarily from the systematically underpredicted PES curvature, which derives from the biased sampling of near-equilibrium atomic arrangements in uMLIP pre-training datasets. Our findings suggest that a considerable fraction of uMLIP errors are highly systematic, and can therefore be efficiently corrected. We argue for the importance of a comprehensive materials dataset with improved PES sampling for next-generation foundational MLIPs.

36 MATERIALS SCIENCE↗
DOE OSTI2021

Attention Enabled Multi-Agent DRL for Decentralized Volt-VAR Control of Active Distribution System Using PV Inverters and SVCs

This paper proposes attention enabled multi-agent deep reinforcement learning (MADRL) framework for active distribution network decentralized Volt-VAR control. Using the unsupervised clustering, the whole distribution system can be decomposed into several sub-networks according to the voltage and reactive power sensitivity relationships. Then, the distributed control problem of each sub-network is modeled as Markov games and solved by the improved MADRL algorithm, where each sub-network is modeled as an adaptive agent. An attention mechanism is developed to help each agent focus on specific information that is mostly related to the reward. All agents are centrally trained offline to learn the optimal coordinated Volt-VAR control strategy and executed in a decentralized manner to make online decisions with only local information. Compared with other distributed control approaches, the proposed method can effectively deal with uncertainties, achieve fast decision makings, and significantly reduce the communication requirements. Comparison results with model-based and other data-driven methods on IEEE 33-bus and 123-bus systems demonstrate the benefits of the proposed approach.

distribution network↗
NASA NTRSDate not supplied

AMSR-2 Daily Snow Depth Data Product Using a Neural Network Algorithm Trained by Collocated ICESat-2 Measurements

By using diffusion theory and Monte Carlo lidar radiative transfer simulations, Hu 1 et al. (2022b) has derived snow depth from the first-, second- and third-order moments of the lidar backscattering pathlength distribution. Lu 2 et al. (2022,2024) calculated the snow depth by applying the methods to the satellite ICESat-2 lidar measurements over the Arctic sea ice, as well as land surfaces of Northern Hemisphere. In this paper, a neural network (NN) algorithm, employing several channels from AMSR-2 and the humidity vertical profiles GMAO GEOS-IT, is trained to determine snow depth identified by time and geolocation matched 2019 ICESat-2 snow-depth data during winter months over the Arctic sea ice. The trained NN snow-depth was applied to 2014-2020 AMSR-2 clear pixel data, although the algorithms perform reasonably well in thinner clouds. This paper used AMSR-2 data, a passive microwave instrument to generate a wide range of snow depth data, covering extensive spatial areas in the cross-orbit direction.

Neural Network↗
DOE OSTI2021

Energy Budget Constraints on the Time History of Aerosol Forcing and Climate Sensitivity

Abstract An observationally constrained time series of historical aerosol effective radiative forcing (ERF) from 1750 to 2019 is developed in this study. We find that the time history of aerosol ERFs diagnosed in CMIP6 models exhibits considerable variation and explore how the time history of aerosol forcing influences the probability distributions of present‐day aerosol forcing and emergent metrics such as climate sensitivity. Using a simple energy balance model, trained on CMIP6 climate models and constrained by observed near‐surface warming and ocean heat uptake, we derive estimates for the historical aerosol forcing. We find 2005–2014 mean aerosol ERF to be −1.1 (−1.8 to −0.5) W m −2 relative to 1750. Assuming recently published historical emissions from fossil fuel and industrial sectors and biomass burning emissions from SSP2‐4.5, aerosol ERF in 2019 is −0.9 (−1.5 to −0.4) W m −2 . There is a modest recovery in aerosol forcing (+0.025 W m −2 decade −1 ) between 1980 and 2014. This analysis also gives a 5%–95% range of equilibrium climate sensitivity of 1.8°C –5.1°C (best estimate 3.1°C) with a transient climate response of 1.2°C –2.6°C (best estimate 1.8°C).

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Neural network based analysis of multimodal bond distributions using extended x-ray absorption fine structure spectra

Knowledge of the local coordination environment around atomic species in functional materials is critical for understanding their mechanisms of operation. Heterogeneous mixtures of metal complexes are ubiquitous in catalysts, ionic liquids, molten salts, biological enzymes, and geochemical systems, among many others. Extracting information from ensemble-average measurements about the structural and compositional descriptors of each type of coordination complex comprising the mixture is not generally possible, especially when they possess multimodal bond-length distributions. Here, we developed a method that enables the mapping of an x-ray absorption spectrum on the radial distribution function describing the average environment of the metal ions. The supervised neural network based method utilizes an objective training set, for which the choice of the local structural motifs is completely agnostic to the theoretically expected structure and dynamics of the modeled system. The method was validated using first-principles modeling of structural dynamics of nickel complexation in molten salts, and it applies to a large class of heterogeneous systems, including those studied under in situ and operando conditions.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2020

DEEP CELLULAR RECURRENT NEURAL ARCHITECTURE FOR EFFICIENT MULTIDIMENSIONAL TIME-SERIES DATA PROCESSING

Efficient processing of time series data is a fundamental yet challenging problem in pattern recognition. Though recent developments in machine learning and deep learning have enabled remarkable improvements in processing large scale datasets in many application domains, most are designed and regulated to handle inputs that are static in time. Many real-world data, such as in biomedical, surveillance and security, financial, manufacturing and engineering applications, are rarely static in time, and demand models able to recognize patterns in both space and time. Current machine learning (ML) and deep learning (DL) models adapted for time series processing tend to grow in complexity and size to accommodate the additional dimensionality of time. Specifically, the biologically inspired learning based models known as artificial neural networks that have shown extraordinary success in pattern recognition, tend to grow prohibitively large and cumbersome in the presence of large scale multi-dimensional time series biomedical data such as EEG. Consequently, this work aims to develop representative ML and DL models for robust and efficient large scale time series processing. First, we design a novel ML pipeline with efficient feature engineering to process a large scale multi-channel scalp EEG dataset for automated detection of epileptic seizures. With the use of a sophisticated yet computationally efficient time-frequency analysis technique known as harmonic wavelet packet transform and an efficient self-similarity computation based on fractal dimension, we achieve state-of-the-art performance for automated seizure detection in EEG data. Subsequently, we investigate the development of a novel efficient deep recurrent learning model for large scale time series processing. For this, we first study the functionality and training of a biologically inspired neural network architecture known as cellular simultaneous recurrent neural network (CSRN). We obtain a generalization of this network for multiple topological image processing tasks and investigate the learning efficacy of the complex cellular architecture using several state-of-the?art training methods. Finally, we develop a novel deep cellular recurrent neural network (CDRNN) architecture based on the biologically inspired distributed processing used in CSRN for processing time series data. The proposed DCRNN leverages the cellular recurrent architecture to promote extensive weight sharing and efficient, individualized, synchronous processing of multi-source time series data. Experiments on a large scale multi-channel scalp EEG, and a machine fault detection dataset show that the proposed DCRNN offers state-of-the-art recognition performance while using substantially fewer trainable recurrent units.

Vidyaratne, Lasitha S.↗
NASA NTRS2001

Ar Atmosphere: Implications for Structure and Composition of Mercury's Crust

We examine the possibilities of sustaining an argon atmosphere by diffusion from the upper 10 km of crust, and alternatively by effusion from a molten or previously molten area at great depth . Ar-40 in the atmospheres of the planets is a measure of potassium abundance in the interiors since Ar-40 is a product of radiogenic decay of K-40 by electron capture with the subsequent emission of a 1.46 eV gamma-ray. Although the Ar-40 in the earth's atmosphere is expected to have accumulated since the late bombardment, Ar-40 in surface-bounded exospheres is eroded quickly by photoionization and electron impact ionization. Thus, the argon content in the exospheres of the Moon, Mercury and probably Europa is representative of current effusion rather than accumulation over the lifetime of the body. Argon content will be a function of K content, temperature, grain size distribution, connected pore volume and possible seismic activity. Although Mercury and the Moon differ in many details, we can train the solutions to diffusion equations to predict the average lunar atmosphere. Then these parameters can be varied for Hermean conditions. Assuming a lunar crustal potassium abundance of 300 ppm, the observed argon atmosphere requires equilibrium between the argon production in the upper 9 Km of the moon (1.135 x 10(exp -3) cm(exp -3) s(exp -1)) and its loss. Hodges et al. conclude that this loss rate and the observed time variability requires argon release through seismic activity, tapping a deep argon source. An important observation is that the extreme surface of the Moon is enhanced in argon rather than depleted, as one would expect from outgassing of radiogenic argon. Manka and Michel concluded that ion implantation explains the surface enhancement of Ar-40. About half of the argon ions produced in the lunar atmosphere would return to the surface, where they would become embedded in the rocks. Similarly, at Mercury we expect the surface rocks to be enhanced in Ar-40 wherever the magnetosphere has been open over time. Thus the measurement of surface composition will reveal the long-term effects of solar wind-magnetosphere interaction. Additional information is contained in the original extended abstract.

Killen, R. M.↗
NASA NTRS2005

Conflict Resolution Performance in an Experimental Study of En Route Free Maneuvering Operations

NASA has developed a far-term air traffic management concept, termed Distributed Air/Ground Traffic Management (DAG-TM). One component of DAG-TM, En Route Free Maneuvering, allows properly trained flight crews of equipped autonomous aircraft to assume responsibility for separation from other autonomous aircraft and from Instrument Flight Rules (IFR) aircraft. Ground-based air traffic controllers continue to separate IFR traffic and issue flow management constraints to all aircraft. To examine En Route Free Maneuvering operations, a joint human-in-the-loop experiment was conducted in summer 2004 at the NASA Ames and Langley Research Centers. Test subject pilots used desktop flight simulators to resolve traffic conflicts and adhere to air traffic flow constraints issued by subject controllers. The experimental airspace integrated both autonomous and IFR aircraft at varying traffic densities. This paper presents a subset of the En Route Free Maneuvering experimental results, focusing on airborne and ground-based conflict resolution, and the effects of increased traffic levels on the ability of pilots and air traffic controllers to perform this task. The results show that, in general, increases in autonomous traffic do not significantly impact conflict resolution performance. In addition, pilot acceptability of autonomous operations remains high throughout the range of traffic densities studied. Together with previously reported findings, these results continue to support the feasibility of the En Route Free Maneuvering component of DAG-TM.

Doble, Nathan A.↗