Search NASA⌕ Search

SEARCH · Search NASA

Results for “training efficiency”

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 127 records · Page 7
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Adaptive Data-Driven Deep-Learning Surrogate Model for Frontal Polymerization in Dicyclopentadiene

Frontal polymerization (FP) is a self-sustaining curing process that enables rapid and energy-efficient manufacturing of thermoset polymers and composites. Computational methods conventionally used to simulate the FP process are time-consuming, and repeating simulations are required for sensitivity analysis, uncertainty quantification, or optimization of the manufacturing process. Here, in this work, we develop an adaptive surrogate deep-learning model for FP of dicyclopentadiene (DCPD), which predicts the evolution of temperature and degree of cure orders of magnitude faster than the finite-element method (FEM). The adaptive algorithm provides a strategy to select training samples efficiently and save computational costs by reducing the redundancy of FEM-based training samples. The adaptive algorithm calculates the residual error of the FP governing equations using automatic differentiation of the deep neural network. A probability density function expressed in terms of the residual error is used to select training samples from the Sobol sequence space. The temperature and degree of cure evolution of each training sample are obtained by a 2D FEM simulation. The adaptive method is more efficient and has a better prediction accuracy than the random sampling method. With the well-trained surrogate neural network, the FP characteristics (front speed, shape, and temperature) can be extracted quickly from the predicted temperature and degree-of-cure fields.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2025

Teacher-student training improves the accuracy and efficiency of machine learning interatomic potentials

Machine learning interatomic potentials (MLIPs) are revolutionizing the field of molecular dynamics (MD) simulations. Recent MLIPs have tended towards more complex architectures trained on larger datasets. The resulting increase in computational and memory costs may prohibit the application of these MLIPs to perform large-scale MD simulations. Herein, we present a teacher-student training framework in which the latent knowledge from the teacher (atomic energies) is used to augment the students' training. We show that the light-weight student MLIPs have faster MD speeds at a fraction of the memory footprint compared to the teacher models. Remarkably, the student models can even surpass the accuracy of the teachers, even though both are trained on the same quantum chemistry dataset. Our work highlights a practical method for MLIPs to reduce the resources required for large-scale MD simulations.

36 MATERIALS SCIENCE↗
DOE OSTI2025

FedEFsz: Fair Cross-Silo Federated Learning System With Error-Bounded Lossy Compression

Cross-Silo federated learning systems have been identified as an efficient approach to scaling DNN training across geographically-distributed data silos to preserve the privacy of the training data. Communication efficiency and fairness are two major issues that need to be both satisfied when federated learning systems are deployed in practice. Simultaneously guaranteeing both of them, however, is exceptionally difficult because simply combining communication reduction and fairness optimization approaches often causes non-converged training or drastic accuracy degradation. Here, to bridge this gap, we propose FedEFsz. On the one hand, it integrates the state-of-the-art error-bounded lossy compressor SZ3 into cross-silo federated learning systems to significantly reduce communication traffic during the training. On the other hand, it achieves a high fairness (i.e., rather consistent model accuracy and performance across different clients) through a carefully designed heuristic algorithm that can tune the error-bound of SZ3 for different clients during the training. Extensive experimental results based on a GPU cluster with 65 GPU cards show that FedEFsz improves the fairness across different benchmarks by up to 60.88% and meanwhile reduces the communication traffic by up to 315×.

Cross-Silo Federated Learning Systems↗
DOE OSTI2024

Decode the Workload: Training Deep Learning Models for Efficient Compute Cluster Representation

Monitoring the status of a high throughput computing cluster running computationally intensive production jobs is a crucial yet challenging system administration task due to the complexity of such systems. To this end, we train autoencoders using the Linux kernel CPU metrics of the cluster. Additionally, we explore assisting these models with graph neural networks to share information across threads within a compute node. The models are compared in terms of their ability to: 1) Produce a compressed latent representation that captures the salient features of the input, 2) Detect anomalous activity, and 3) Make distinction between different kinds of jobs run at Jefferson Lab. The goal is to have a robust encoder whose compressed embeddings are used for several downstream tasks. We extend this study further by deploying these models in a human-in-the-loop production-based setting for the anomaly detection task and discuss the associated implementation aspects such as continual learning and the criterion to generate alarms. This study represents a first step in the endeavor towards building self-supervised large-scale foundation models for computing centers.

Mohammed, Ahmed↗
DOE OSTI2025

Scalable training of trustworthy and energy-efficient predictive graph foundation models for atomistic materials modeling: a case study with HydraGNN

We present our work on developing and training scalable, trustworthy, and energy-efficient predictive graph foundation models (GFMs) using HydraGNN, a multi-headed graph convolutional neural network architecture. HydraGNN expands the boundaries of graph neural network (GNN) computations in both training scale and data diversity. It abstracts over message passing algorithms, allowing both reproduction of and comparison across algorithmic innovations that define nearest-neighbor convolution in GNNs. This work discusses a series of optimizations that have allowed scaling up the GFMs training to tens of thousands of GPUs on datasets consisting of hundreds of millions of graphs. Our GFMs use multitask learning (MTL) to simultaneously learn graph-level and node-level properties of atomistic structures, such as energy and atomic forces. Using over 154 million atomistic structures for training, we illustrate the performance of our approach along with the lessons learned on two state-of-the-art US Department of Energy (US-DOE) supercomputers, namely the Perlmutter petascale system at the National Energy Research Scientific Computing Center and the Frontier exascale system at Oak Ridge Leadership Computing Facility. The HydraGNN architecture enables the GFM to achieve near-linear strong scaling performance using more than 2000 GPUs on Perlmutter and 16,000 GPUs on Frontier.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS2023

A Global Land Cover Training Dataset From 1984 to 2020

State-of-the-art cloud computing platforms such as Google Earth Engine (GEE) enable regional-to-global land cover and land cover change mapping with machine learning algorithms. However, collection of high-quality training data, which is necessary for accurate land cover mapping, remains costly and labor-intensive. To address this need, we created a global database of nearly 2 million training units spanning the period from 1984 to 2020 for seven primary and nine secondary land cover classes. Our training data collection approach leveraged GEE and machine learning algorithms to ensure data quality and biogeographic representation. We sampled the spectral-temporal feature space from Landsat imagery to efficiently allocate training data across global ecoregions and incorporated publicly available and collaborator-provided datasets to our database. To reflect the underlying regional class distribution and post-disturbance landscapes, we strategically augmented the database. We used a machine learning-based cross-validation procedure to remove potentially mis-labeled training units. Our training database is relevant for a wide array of studies such as land cover change, agriculture, forestry, hydrology, urban development, among many others.

Radost Stanimirova↗
NASA NTRS2011

Houston, We Have a Problem Solving Model for Training

In late 2006, the Mission Operations Directorate (MOD) at NASA began looking at ways to make training more efficient for the flight controllers who support the International Space Station. The average certification times for flight controllers spanned from 18 months to three years and the MOD, responsible for technical training, was eager to develop creative solutions that would reduce the time to 12 months. Additionally, previously trained flight controllers sometimes participated in more than 50 very costly, eight-hour integrated simulations before becoming certified. New trainees needed to gain proficiency with far fewer lessons and training simulations than their predecessors. This poster presentation reviews the approach and the process that is currently in development to accomplish this goal.

Schmidt, Lacey↗
DOE OSTI2022

BUTTER-E - Energy Consumption Data for the BUTTER Empirical Deep Learning Dataset

The BUTTER-E - Energy Consumption Data for the BUTTER Empirical Deep Learning Dataset adds node-level energy consumption data from watt-meters to the primary sweep of the BUTTER - Empirical Deep Learning Dataset. This dataset contains energy consumption and performance data from 63,527 individual experimental runs spanning 30,582 distinct configurations: 13 datasets, 20 sizes (number of trainable parameters), 8 network "shapes", and 14 depths on both CPU and GPU hardware collected using node-level watt-meters. This dataset reveals the complex relationship between dataset size, network structure, and energy use, and highlights the impact of cache effects. BUTTER-E is intended to be joined with the BUTTER dataset (see "BUTTER - Empirical Deep Learning Dataset on OEDI" resource below) which characterizes the performance of 483k distinct fully connected neural networks but does not include energy measurements.

Array↗
NASA NTRS2018

Next Gen One Portal Usability Evaluation

Each exercise device on the International Space Station (ISS) has a unique, customized software system interface with unique layouts / hierarchy, and operational principles that require significant crew training. Furthermore, the software programs are not adaptable and provide no real-time feedback or motivation to enhance the exercise experience and/or prevent injuries. Additionally, the graphical user interfaces (GUI) of these systems present information through multiple layers resulting in difficulty navigating to the desired screens and functions. These limitations of current exercise device GUI's lead to increased crew time spent on initiating, loading, performing exercises, logging data and exiting the system. To address these limitations a Next Generation One Portal (NextGen One Portal) Crew Countermeasure System (CMS) was developed, which utilizes the latest industry guidelines in GUI designs to provide an intuitive ease of use approach (i.e., 80% of the functionality gained within 5-10 minutes of initial use without/limited formal training required). This is accomplished by providing a consistent interface using common software to reduce crew training, increase efficiency & user satisfaction while also reducing development & maintenance costs. Results from the usability evaluations showed the NextGen One Portal UI having greater efficiency, learnability, memorability, usability and overall user experience than the current Advanced Resistive Exercise Device (ARED) UI used by astronauts on ISS. Specifically, the design of the One-Portal UI as an app interface similar to those found on the Apple and Google's App Store, assisted many of the participants in grasping the concepts of the interface with minimum training. Although the NextGen One-Portal UI was shown to be an overall better interface, observations by the test facilitators noted specific exercise tasks appeared to have a significant impact on the NextGen One-Portal UI efficiency. Future updates to the NextGen One Portal UI will address these inefficiencies.

Cross, E. V., III↗
NASA NTRSDate not supplied

Real-Time Pilot Inputs for Aircraft Dynamic Modeling

Real-time dynamic modeling and pilot displays are used to enable pilots to intuitively discover effective multi-axis inputs for aircraft dynamic modeling during flight tests. An F-16 nonlinear simulation flown by a pilot using a joystick and a laptop computer with real-time displays of relevant information is used to demonstrate the approach. Results show that pilots can use the real-time display to discover and implement effective and efficient multi-axis inputs for accurate dynamic modeling in flight. The real-time dynamic modeling results driving the pilot display can be used to certify that the acquired flight data are adequate for accurate dynamic modeling, thus avoiding postflight data analysis and modeling by an analyst to make that determination. Implications for pilot training and efficient flight testing are discussed.

Pilot Inputs↗
DOE OSTI2021

Adaptive pruning-based optimization of parameterized quantum circuits

Abstract Variational hybrid quantum–classical algorithms are powerful tools to maximize the use of noisy intermediate-scale quantum devices. While past studies have developed powerful and expressive ansatze, their near-term applications have been limited by the difficulty of optimizing in the vast parameter space. In this work, we propose a heuristic optimization strategy for such ansatze used in variational quantum algorithms, which we call ‘parameter-efficient circuit training (PECT)’. Instead of optimizing all of the ansatz parameters at once, PECT launches a sequence of variational algorithms, in which each iteration of the algorithm activates and optimizes a subset of the total parameter set. To update the parameter subset between iterations, we adapt the Dynamic Sparse Reparameterization scheme which was originally proposed for training deep convolutional neural networks. We demonstrate PECT for the Variational Quantum Eigensolver, in which we benchmark unitary coupled-cluster ansatze including UCCSD and k -UpCCGSD, as well as the Low-Depth Circuit Ansatz (LDCA), to estimate ground state energies of molecular systems. We additionally use a layerwise variant of PECT to optimize a hardware-efficient circuit for the Sycamore processor to estimate the ground state energy densities of the one-dimensional Fermi-Hubbard model. From our numerical data, we find that PECT can enable optimizations of certain ansatze that were previously difficult to converge and more generally can improve the performance of variational algorithms by reducing the optimization runtime and/or the depth of circuits that encode the solution candidate(s).

Physics↗
NASA NTRSDate not supplied

Reinforcement Learning for Spacecraft Navigation & Environment Characterization in the Planar-Restricted Two-Body Problem

As science, exploration, and commercial space missions become increasingly complex, so does the need for efficient, autonomous, and integrated spacecraft navigation and operations techniques. Key operational functions, including data collection and transmission, environment characterization, systems constraints, human factors, and navigation, often are intertwined and conflicted. Deep Reinforcement Learning (DRL) offers a framework for addressing integrated spacecraft navigation and planning in an uncertain dynamical environment. The goal of this study is to evaluate the utility of DRL for integrated spacecraft navigation and planning. This is achieved by developing a simple environmental characterization training environment in the Planar-Restricted 2-Body Problem (PR2BP), establishing benchmarks and heuristic baselines, and designing a previously unstudied Markov Decision Process (MDP) formulation. This MDP formulation enables the spacecraft DRL agents to appropriately balance navigation and actuation capabilities. The resulting DRL-derived policy exceeds a random or untrained policy and meets or exceeds the level of performance of a heuristic without actuation. In the process, valuable intuition is gained about the problem with insight into how DRL methods could scale to increasingly more realistic scenarios, including net-work design and training architectures, efficient state space representations, and methods for encouraging exploration in a parametric action space, among others.

navigation↗
NASA NTRSDate not supplied

Materials Informatics at NASA GRC: Machine Learning Surrogate Modeling, Data Management, and Integrated Toolsets for Establishing/Maintaining the Digital Thread

Integrated Computational Materials Engineering (ICME) has recently received widespread attention due to its promises in reducing dependence on physical testing for engineering design by relying on simulation, reducing both time and cost to market for various applications. ICME however requires validated multiscale material models, which heavily depend on available test data with full material and test pedigree, including material processing, test and measurement equipment, raw data collection, and analysis methodology and results that is findable and usable, along with integrated, efficient toolsets for effectively passing information across various length and time scales across such models. At the NASA Glenn Research Center under the Transformational Tools and Technologies Project, significant recent efforts have been directed towards establishing the required cyberinfrastructure to enable optimized ICME processes and the design of “fit-for-purpose” materials to achieve the goals outlined in the NASA Vision 2040 report. Such efforts include development of multiscale physics-based material models, which can be used to train highly efficient surrogate machine learning models, development of best practices and infrastructure for effective, traceable materials information management, and development of toolsets that integrate with physics-based codes, machine learning models, and an information management system to enable high throughput of materials data collection and analysis, establishment of digital twins and the digital thread, and automation of the ICME design process for material optimization.

Machine Learning↗
DOE OSTI2023

$\text{GPLaSDI}$: Gaussian Process-based interpretable Latent Space Dynamics Identification through deep autoencoder

Numerically solving partial differential equations (PDEs) can be challenging and computationally expensive. This has led to the development of reduced-order models (ROMs) that are accurate but faster than full order models (FOMs). Recently, machine learning advances have enabled the creation of non-linear projection methods, such as Latent Space Dynamics Identification (LaSDI). LaSDI maps full-order PDE solutions to a latent space using autoencoders and learns the system of ODEs governing the latent space dynamics. By interpolating and solving the ODE system in the reduced latent space, fast and accurate ROM predictions can be made by feeding the predicted latent space dynamics into the decoder. In this paper, we introduce GPLaSDI, a novel LaSDI-based framework that relies on Gaussian process (GP) for latent space ODE interpolations. Using GPs offers two significant advantages. First, it enables the quantification of uncertainty over the ROM predictions. Second, leveraging this prediction uncertainty allows for efficient adaptive training through a greedy selection of additional training data points. This approach does not require prior knowledge of the underlying PDEs. Consequently, GPLaSDI is inherently non-intrusive and can be applied to problems without a known PDE or its residual. Here we demonstrate the effectiveness of our approach on the Burgers equation, Vlasov equation for plasma physics, and a rising thermal bubble problem. Our proposed method achieves between 200 and 100,000 times speed-up, with up to 7% relative error.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

A data-driven multiscale model for reactive wetting simulations

Here, we describe a data-driven, multiscale technique to model reactive wetting of a silver–aluminum alloy on a Kovar™ (Fe-Ni-Co alloy) surface. We employ molecular dynamics simulations to elucidate the dependence of surface tension and wetting angle on the drop’s composition and temperature. A design of computational experiments is used to efficiently generate training data of surface tension and wetting angle from a limited number of molecular dynamics simulations. The simulation results are used to parameterize models of the material’s wetting properties and compute the uncertainty in the models due to limited data. The data-driven models are incorporated into an engineering-scale (continuum) model of a silver–aluminum sessile drop on a Kovar™ substrate. Model predictions of the wetting angle are compared with experiments of pure silver spreading on Kovar™ to quantify the model-form errors introduced by the limited training data versus the simplifications inherent in the molecular dynamics simulations. The paper presents innovations in the determination of “convergence” of noisy MD simulations before they are used to extract the wetting angle and surface tension, and the construction of their models which approximate physio-chemical processes that are left unresolved by the engineering-scale model. Together, these constitute a multiscale approach that integrates molecular-scale information into continuum scale models.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2021

Community-Driven Methods for Open and Reproducible Software Tools for Analyzing Datasets from Atom Probe Microscopy

Atom probe tomography, and related methods, probe the three-dimensional architecture of a material. The software tools that microscopists use, and how these tools are connected into workflows, makes a substantial contribution to the accuracy and precision of such a material characterization experiment. Typically, we adapt methods from other communities like mathematics, data science, computational geometry, artificial intelligence, or scientific computing. We also realize that improving on research data management is a challenge when it comes to align with the FAIR data stewardship principles. Faced with this global challenge, we are convinced that collaborating is useful. Here, we report the results and challenges with an inter-laboratory call for developing test cases for several types of atom probe software tools. The results support why defining detailed recipes of software workflows and sharing these recipes is necessary and rewarding: Open source tools and (meta)data exchange can help to make our day-to-day data processing tasks become more efficient, the training of new users and knowledge transfer become easier, and assist us with automated quantification of uncertainties to gain access to substantiated results.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Accelerating the design of lattice structures using machine learning

Lattices remain an attractive class of structures due to their design versatility; however, rapidly designing lattice structures with tailored or optimal mechanical properties remains a significant challenge. With each added design variable, the design space quickly becomes intractable. To address this challenge, research efforts have sought to combine computational approaches with machine learning (ML)-based approaches to reduce the computational cost of the design process and accelerate mechanical design. While these efforts have made substantial progress, significant challenges remain in (1) building and interpreting the ML-based surrogate models and (2) iteratively and efficiently curating training datasets for optimization tasks. Here, we address the first challenge by combining ML-based surrogate modeling and Shapley additive explanation (SHAP) analysis to interpret the impact of each design variable. We find that our ML-based surrogate models achieve excellent prediction capabilities (R 2 > 0.95) and SHAP values aid in uncovering design variables influencing performance. We address the second challenge by utilizing active learning-based methods, such as Bayesian optimization, to explore the design space and report a 5 × reduction in simulations relative to grid-based search. Collectively, these results underscore the value of building intelligent design systems that leverage ML-based methods for uncovering key design variables and accelerating design.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Ultrafast jet classification at the HL-LHC

Abstract Three machine learning models are used to perform jet origin classification. These models are optimized for deployment on a field-programmable gate array device. In this context, we demonstrate how latency and resource consumption scale with the input size and choice of algorithm. Moreover, the models proposed here are designed to work on the type of data and under the foreseen conditions at the CERN large hadron collider during its high-luminosity phase. Through quantization-aware training and efficient synthetization for a specific field programmable gate array, we show that O ( 100 ) ns inference of complex architectures such as Deep Sets and Interaction Networks is feasible at a relatively low computational resource cost.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗