Search NASA⌕ Search

SEARCH · Search NASA

Results for “compressed 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 37 records · Page 2
All sourcesNASA NTRSDOE OSTI
DOE OSTI2023

Stability of the modulator in a plasma-modulated plasma accelerator

We explore the regime of operation of the modulator stage of a recently proposed laser-plasma accelerator scheme [Phys. Rev. Lett. 127, 184801 (2021)], dubbed the plasma-modulated plasma accelerator (P-MoPA). The P-MoPA scheme offers a potential route to high-repetition-rate, GeV-scale plasma accelerators driven by picosecond-duration laser pulses from, for example, kilohertz thin-disk lasers. The first stage of the P-MoPA scheme is a plasma modulator in which a long, high-energy “drive” pulse is spectrally modulated by copropagating in a plasma channel with the low-amplitude plasma wave driven by a short, low-energy “seed” pulse. The spectrally modulated drive pulse is converted to a train of short pulses, by introducing dispersion, which can resonantly drive a large wakefield in a subsequent accelerator stage with the same on-axis plasma density as the modulator. In this paper we derive the 3D analytic theory for the evolution of the drive pulse in the plasma modulator and show that the spectral modulation is independent of transverse coordinate, which is ideal for compression into a pulse train. We then identify a transverse mode instability (TMI), similar to the TMI observed in optical fiber lasers, which sets limits on the energy of the drive pulse for a given set of laser-plasma parameters. We compare this analytic theory with particle-in-cell (PIC) simulations and find that even higher energy drive pulses can be modulated than those demonstrated in the original proposal.

43 PARTICLE ACCELERATORS↗
DOE OSTI2026

Dense autoencoders, clustering techniques, and semi-supervised learning for HPGe $γ$-spectra

Classifying high-resolution gamma spectra by their isotopic content is an essential task in nuclear forensics and other applications. Traditional analysis methods are often time-intensive, but machine learning (ML) may help analysts quickly process many spectra. Such methods tend to rely on abundant, well-labeled data for training. Historical gamma data exists in various fields but is not uniformly useful for supervised ML due to inconsistent labeling. Here, to address some of these challenges, we present a method to classify and organize unlabeled data from high-purity germanium detectors using an autoencoding neural network (autoencoder). We trained dense autoencoders to compress gamma data into latent representations that enable efficient data characterization. By clustering the encoded spectra or lower-dimensional mappings of them, we identified and removed portions of over-abundant data categories, resulting in a more balanced dataset and improved autoencoder performance. This encoding and clustering pipeline also enabled the organization of spectra into self-consistent categories. Finally, we found that encoded representations showed potential as inputs for semi-supervised learning of nuclide identification (NID) labels, achieving an average F1 score of 0.85 ± 0.03 when mapping encodings to a set of 65 isotope labels.

Autoencoders↗
DOE OSTI2025

Normalizing flows for high-dimensional detector simulations

Whenever invertible generative networks are needed for LHC physics, normalizing flows show excellent performance. In this work, we investigate their performance for fast calorimeter shower simulations with increasing phase space dimension. We use fast and expressive coupling spline transformations applied to the CaloChallenge datasets. In addition to the base flow architecture we also employ a VAE to compress the dimensionality and train a generative network in the latent space. We evaluate our networks on several metrics, including high-level features, classifiers, and generation timing. Our findings demonstrate that invertible neural networks have competitive performance when compared to autoregressive flows, while being substantially faster during generation.

Ernst, Florian↗
DOE OSTI2026

3D Gaussian Splatting for Volume Compression

This codebase uses machine learning to train a collection of 3D Gaussian distributions to approximate scientific volume data. Because this collection uses less memory than the original dataset, it can be used as a compressed model of the original data for applications such as visualization.

Dyken, Landon↗
DOE OSTI2021

Controllable X-ray Pulse Trains from Enhanced Self-Amplified Spontaneous Emission

We report the demonstration of optical compression of an electron beam and the production of controllable trains of femtosecond, soft x-ray pulses with the Linac Coherent Light Source (LCLS) freeelectron laser (FEL). This is achieved by enhanced self-amplified spontaneous emission with a 2 μm laser and a dechirper device. Optical compression was achieved by modulating the energy of an electron beam with the laser and then compressing with a chicane, resulting in high current spikes on the beam which we observe to lase. A dechirper was then used to selectively control the lasing region of the electron beam. Field autocorrelation measurements indicate a train of pulses, and we find that the number of pulses within the train can be controlled (from 1 to 5 pulses) by varying the dechirper position and undulator taper. These results are a step toward attosecond spectroscopy with x-ray FELs as well as future FEL schemes relying on optical compression of an electron beam.

Duris, Joseph P↗
DOE OSTI2024

Liquid Piston with Spray Cooling Near-Isothermal Compressor

The goal of this project was to prototype and characterize the performance of a liquid-piston spray-cooled gas compressor. The working principle of the compressor enables optimized high-efficiency operation over a very wide range of operating conditions, unlike conventional compressors that are optimized for a narrow range of operating conditions. The compressor technology is suitable for many applications, such as gas pipeline transport, gas storage, and commercial and residential heat pumps. Both physical testing and computational fluid dynamics (CFD) modeling of the processes using the Oak Ridge National Laboratory high-performance computing center were completed. The experimental and CFD studies focused on a near-isothermal liquid-piston compressor (LPC) that uses propylene glycol to compress CO 2 . The first prototype demonstrated isothermal operation during several sequentially executed cycles of CO 2 compression and raised the temperature of the compressed CO 2 by only 2 K, compared with approximately 6 K when the gas was compressed non-isothermally. Isothermal operation was demonstrated at CO 2 flow rates of up to 2 L/min. The second prototype was designed with two compression chambers to allow continuous flow of high-pressure CO 2 . However, the design of the valve train to direct flow between the compression chambers was not sufficient to allow demonstration of CO 2 compression. Numerical simulations of the LPC in which the compression chamber was filled with propylene glycol injected from the bottom inlet were performed using large eddy simulation (LES) with the wall-adapting local eddy-viscosity subgrid-scale model coupled with the multiphase volume of fluid (VOF) model to simulate the transient interface between gas and liquid and to capture the heat and mass transfers within the compression chamber. In this effort, the effects of boundary conditions applied to the LES-VOF calculations (i.e., no wall, an adiabatic wall, and a wall with a heat flux subscribed) on the overall pressure and temperature of the CO 2 gas as well as the transient evolution of flow and heat transfer within the compression chamber were investigated and are discussed in this report. The LES calculation with no wall showed no dynamical flow patterns, and the volume-averaged temperature of CO 2 increased from 305 to 392.7 K, whereas LES calculations with a constant wall temperature or a wall heat flux had similar increases of CO 2 temperatures. The results of the LES simulation using a wall heat flux showed different stages in the compression process and revealed dynamical formation and interaction of CO 2 gas layers and circulation flow patterns within the chamber that contributed to the overall heat transfer between the solid wall, gas, and liquid surface in the compressor. Though an industrial partnership for commercializing the compressor was not secured, the technology was attractive for an industrial partner to use in two research proposals in response to US Department of Energy funding opportunity announcements.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2025

FedCSpc: A Cross-Silo Federated Learning System With Error-Bounded Lossy Parameter Compression

Cross-Silo federated learning is widely used for scaling deep neural network (DNN) training over data silos from different locations worldwide while guaranteeing data privacy. Communication has been identified as the main bottleneck when training large-scale models due to large-volume model parameters and gradient transmission across public networks with limited bandwidth. Most previous works focus on gradient compression, while limited work tries to compress parameters that can not be ignored and extremely affect communication performance during the training. Here, to bridge this gap, we propose FedCSpc: an efficient cross-silo federated learning system with an XAI-driven adaptive parameter compression strategy for large-scale model training. Our work substantially differs from existing gradient compression techniques due to the distinct data features of gradient and parameter. The key contributions of this paper are fourfold. (1) Our designed FedCSpc proposes to compress the parameter during the training using the state-of-the-art error-bounded lossy compressor – SZ3. (2) We develop an adaptive compression error bound adjustment algorithm to guarantee the model accuracy effectively. (3) We exploit an efficient approach to utilize the idle CPU resources of clients to compress the parameters. (4) We perform a comprehensive evaluation with a wide range of models and benchmarks on a GPU cluster with 65 GPUs. Results show that FedCSpc can achieve the same model accuracy as FedAvg while reducing the data volume of parameters and gradients in communication by up to 7.39× and 288×, respectively. With 32 clients on a 4 Gb size model, FedCSpc significantly outperforms FedAvg in wall-clock time in the emulated WAN environment (at the bandwidth of 1 Gbps or lower without loss of generality).

SZ3↗
DOE OSTI2023

A MHz X-ray diffraction set-up for dynamic compression experiments in the diamond anvil cell

An experimental platform for dynamic diamond anvil cell (dDAC) research has been developed at the High Energy Density (HED) Instrument at the European X-ray Free Electron Laser (European XFEL). Advantage was taken of the high repetition rate of the European XFEL (up to 4.5 MHz) to collect pulse-resolved MHz X-ray diffraction data from samples as they are dynamically compressed at intermediate strain rates (≤10 3 s −1 ), where up to 352 diffraction images can be collected from a single pulse train. The set-up employs piezo-driven dDACs capable of compressing samples in ≥340 µs, compatible with the maximum length of the pulse train (550 µs). Results from rapid compression experiments on a wide range of sample systems with different X-ray scattering powers are presented. A maximum compression rate of 87 TPa s −1 was observed during the fast compression of Au, while a strain rate of ∼1100 s −1 was achieved during the rapid compression of N 2 at 23 TPa s −1 .

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2023

Out of distribution evaluation for neural image compression

Official code for the NeurIPS 2023 paper "Neural Image Compression: Generalization, Robustness, and Spectral Biases". This code can be used to: - Computing spectral distortion errors of images compressed with traditional codes or neural image compression models (as proposed in the aformentioned paper) - Visualize power spectral densities and fourier heatmaps (as proposed in the aformentioned paper) - Train and test a variety of neural image compression models

Lieberman, KelseyA↗
DOE OSTI2026

Maximizing efficiency of dataset compression for machine learning potentials with information theory

Machine learning interatomic potentials (MLIPs) balance high accuracy and lower costs compared to density functional theory calculations, but their performance often depends on the size and diversity of training datasets. Large datasets improve model accuracy and generalization but are computationally expensive to produce and train on, while smaller datasets risk discarding rare but important atomic environments and compromising MLIP accuracy/reliability. Here, we develop an information-theoretical framework to quantify the efficiency of dataset compression methods and propose an algorithm that maximizes this efficiency. By framing atomistic dataset compression as an instance of the minimum set cover (MSC) problem over atom-centered environments, our method identifies the smallest subset of structures that contains as much information as possible from the original dataset while pruning redundant information. The approach is extensively demonstrated on the GAP-20 and TM23 datasets and validated on 64 varied datasets from the ColabFit repository. Across all cases, MSC consistently retains outliers, preserves dataset diversity, and reproduces the long-tail distributions of forces even at high compression rates, outperforming other subsampling methods. Furthermore, MLIPs trained on MSC-compressed datasets exhibit reduced error for out-of-distribution data even in low-data regimes. We explain these results using an outlier analysis and show that such quantitative conclusions could not be achieved with conventional dimensionality reduction methods. The algorithm is implemented in the open-source QUESTS package and can be used for several tasks in atomistic modeling, from data subsampling, outlier detection, and training improved MLIPs at a lower cost.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Predicting non-linear stress–strain response of mesostructured cellular materials using supervised autoencoder

Recent breakthroughs in advanced manufacturing capabilities have made it possible to design and print sophisticated topologies of cellular structures using diverse engineering materials such as metals, polymers, and ceramics. In these architectured materials, it is often desirable to tailor the mechanical properties by altering the unit cell topology. This necessitates an in-depth understanding of how the topology of the unit cell structure affects the macroscopic behavior of the material in both the linear and the non-linear regimes encountered under large compression. Here, we have developed a machine learning (ML) approach capable of accelerating the prediction of the stress–strain response of a polymer-based cellular structure under uniaxial confined compression. As part of generating the training data for ML, 60,000 mesostructures were generated using a relatively novel approach based on cellular automata, and their corresponding stress–strain responses were obtained from the finite element simulations. Principal component analysis (PCA) was used to reduce the dimensionality of the stress–strain curves. With only 20 principal components, PCA captured 99.89% of the variance in the stress–strain curves while reducing the dimensionality by 5X. ML using supervised autoencoder was able to successfully speed up the prediction of the non-linear stress–strain response of a unit cell by up to 4600X. The proposed method can serve as an efficient data generation tool and a rapid means for predicting the structure–property relationship through accelerated forward modeling of cellular materials under compaction, in cases where the macroscopic stress–strain response is governed by the unit-cell topology.

36 MATERIALS SCIENCE↗
DOE OSTI2020

Machine-learning-assisted insight into spin ice Dy 2 Ti 2 O 7

Complex behavior poses challenges in extracting models from experiment. An example is spin liquid formation in frustrated magnets like Dy 2 Ti 2 O 7 . Understanding has been hindered by issues including disorder, glass formation, and interpretation of scattering data. Here, we use an automated capability to extract model Hamiltonians from data, and to identify different magnetic regimes. This involves training an autoencoder to learn a compressed representation of three-dimensional diffuse scattering, over a wide range of spin Hamiltonians. The autoencoder finds optimal matches according to scattering and heat capacity data and provides confidence intervals. Validation tests indicate that our optimal Hamiltonian accurately predicts temperature and field dependence of both magnetic structure and magnetization, as well as glass formation and irreversibility in Dy 2 Ti 2 O 7 . The autoencoder can also categorize different magnetic behaviors and eliminate background noise and artifacts in raw data. Our methodology is readily applicable to other materials and types of scattering problems.

36 MATERIALS SCIENCE↗
DOE OSTI2023

Physics-Driven Convolutional Autoencoder Approach for CFD Data Compressions: Preprint

With the growing size and complexity of turbulent flow models, data compression approaches are of the utmost importance to analyze, visualize, or restart the simulations. Recently, in-situ autoencoder-based compression approaches have been proposed and shown to be effective at producing reduced representations of turbulent flow data. However, these approaches focus solely on training the model using point-wise sample reconstruction losses that do not take advantage of the physical properties of turbulent flows. In this paper, we show that training autoencoders with additional physics-informed regularizations, e.g., enforcing incompressibility and preserving enstrophy, improves the compression model in three ways: (i) the compressed data better conform to known physics for homogeneous isotropic turbulence without negatively impacting point-wise reconstruction quality, (ii) inspection of the gradients of the trained model uncovers changes to the learned compression mapping that can facilitate the use of explainability techniques, and (iii) as a performance byproduct, training losses are shown to converge up to 12x faster than the baseline model.

auto-encoders↗
DOE OSTI2025

Deciphering the small-angle scattering of polydisperse hard spheres using deep learning

We introduce a deep learning approach for analyzing the scattering function of the polydisperse hard sphere system. We use a variational autoencoder-based neural network to learn the bidirectional mapping between the scattering function and the system parameters, including the volume fraction and polydispersity. Such that the trained model serves both as a generator that produces a scattering function from the system parameters and an inferrer that extracts system parameters from the scattering function. We first generate a scattering dataset by carrying out molecular dynamics simulations of the polydisperse hard spheres modeled by the truncated-shifted Lennard-Jones model, then analyze the scattering function dataset using singular value decomposition to confirm the feasibility of dimensional compression. Then, we split the dataset into training and testing sets and train our neural network on the training set only. Our generator model produces a scattering function with significantly higher accuracy compared to the traditional Percus–Yevick approximation and β correction, and the inferrer model can extract the volume fraction and polydispersity with much higher accuracy than traditional model functions.

Ding, Lijie [ORNL] (ORCID:0000000227454606)↗
DOE OSTI2025

Neural architecture codesign for fast physics applications

We develop a pipeline to streamline neural architecture codesign for physics applications to reduce the need for ML expertise when designing models for novel tasks. Our method employs neural architecture search and network compression in a two-stage approach to discover hardware efficient models. This approach consists of a global search stage that explores a wide range of architectures while considering hardware constraints, followed by a local search stage that fine-tunes and compresses the most promising candidates. We exceed performance on various tasks and show further speedup through model compression techniques such as quantization-aware-training and neural network pruning. We synthesize the optimal models to high level synthesis code for FPGA deployment with the hls4ml library. Additionally, our hierarchical search space provides greater flexibility in optimization, which can easily extend to other tasks and domains. We demonstrate this with two case studies: Bragg peak finding in materials science and jet classification in high energy physics, achieving models with improved accuracy, smaller latencies, or reduced resource utilization relative to the baseline models.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2022

Efficient Clustering of Software Vulnerabilities using Self Organizing Map (SOM)

The common vulnerabilities and exposures (CVE) database was created with a mission to ``identify, define, and catalog publicly disclosed cybersecurity vulnerabilities''. This rich body of information can be used to enable rapid and efficient response to secure and defend cyber operations and protect critical cyber infrastructure. The main goal of this paper is to develop a visual analytics tool to enable deep analysis of CVEs using unsupervised clustering techniques. We enhance our analysis by first mapping CVEs to hierarchical-classes in Common Weakness Enumeration (CWE) using information in the National Vulnerability Database (NVD). Both the mapping and the numerical representation of CVEs are enabled by V2W-BERT, which uses natural language processing of the extensive information in NVD to generate a large tabular database of 137,226 CVE entries from 1999 to 2020, where each CVE is represented by a vector of 768 numerical features. The vectorized data is processed by Self-Organizing Maps (SOM), which is an unsupervised machine learning technique for dimensionality reduction, visual representation and clustering. Using a Torus map of 6417 units, we achieve ~10-fold data compression of ~140k CVEs using SOM. The trained map is further clustered using standard K-means clustering into 138 clusters of CVEs. We conducted a brief investigation of the rich mapping of CVEs to best-matching-units to K-means clusters, as well as CVEs to CWEs. For example, this novel mapping provided insight into the role of CWE-59 and CWE-264 in several CVEs that is otherwise hard to explore in the original data. We conclude that our this novel approach will not only enable deep analysis of the complex relationships between CVEs and CWEs, but also a mechanism to quickly respond to and design mitigation actions for rapidly evolving vulnerabilities that have not been mapped to existing CWEs.

Panchal, Khyati↗
DOE OSTI2023

Differentiable Earth mover’s distance for data compression at the high-luminosity LHC

Abstract The Earth mover’s distance (EMD) is a useful metric for image recognition and classification, but its usual implementations are not differentiable or too slow to be used as a loss function for training other algorithms via gradient descent. In this paper, we train a convolutional neural network (CNN) to learn a differentiable, fast approximation of the EMD and demonstrate that it can be used as a substitute for computing-intensive EMD implementations. We apply this differentiable approximation in the training of an autoencoder-inspired neural network (encoder NN) for data compression at the high-luminosity LHC at CERN The goal of this encoder NN is to compress the data while preserving the information related to the distribution of energy deposits in particle detectors. We demonstrate that the performance of our encoder NN trained using the differentiable EMD CNN surpasses that of training with loss functions based on mean squared error.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2020

In situ compression artifact removal in scientific data using deep transfer learning and experience replay

The massive amount of data produced during simulation on high-performance computers has grown exponentially over the past decade, exacerbating the need for streaming compression and decompression methods for efficient storage and transfer of this data---key to realizing the full potential of large-scale computational science. Lossy compression approaches such as JPEG when applied to scientific simulation data realized as a stream of images can achieve good compression rates but at the cost of introducing compression artifacts and loss of information. This paper develops a unified framework for in situ compression artifact removal in which the fully convolutional neural network architectures are combined with scalable training, transfer learning, and experience replay to achieve superior accuracy and efficiency while significantly decreasing the storage footprint as compared with the traditional optimization-based approaches. We demonstrate the proposed approach and compare it with compressed sensing postprocessing and other baseline deep learning models using climate simulations and nuclear reactor simulations, both of which are driven by hyperbolic partial differential equations. Our approach when applied to remove the compression artifacts on the JPEG-compressed nuclear reactor simulation data (using a transfer-trained model that was pretrained on the climate simulation data and updated incrementally as the nuclear reactor simulation progressed), achieved a significant improvement---mean peak signal-to-noise ratio of 42.438 as compared with 27.725 obtained with the compressed sensing approach.

97 MATHEMATICS AND COMPUTING↗