Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “inference accelerators”

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 145 records · Page 8
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

Thermal Experiments for Fractured Rock Characterization: Theoretical Analysis and Inverse Modeling

Abstract Field‐scale properties of fractured rocks play a crucial role in many subsurface applications, yet methodologies for identification of the statistical parameters of a discrete fracture network (DFN) are scarce. We present an inversion technique to infer two such parameters, fracture density and fractal dimension, from cross‐borehole thermal experiments data. It is based on a particle‐based heat‐transfer model, whose evaluation is accelerated with a deep neural network (DNN) surrogate that is integrated into a grid search. The DNN is trained on a small number of the heat‐transfer model runs and predicts the cumulative density function of the thermal field. The latter is used to compute fine posterior distributions of the (to be estimated) parameters. Our synthetic experiments reveal that fracture density is well constrained by data, while fractal dimension is harder to determine. Adding nonuniform prior information related to the DFN connectivity improves the inference of this parameter.

Zhou, Zitong↗
DOE OSTI2024

Random forest prediction of crystal structure from electron diffraction patterns incorporating multiple scattering

Diffraction is the most common method to solve for unknown or partially known crystal structures. However, it remains a challenge to determine the crystal structure of a new material that may have nanoscale size or heterogeneities. Here, in this study, we train an architecture of hierarchical random forest models capable of predicting the crystal system, space group, and lattice parameters from one or more unknown two-dimensional electron diffraction patterns. Our initial model correctly identifies the crystal system of a simulated electron diffraction pattern from a 20-nm-thick specimen of arbitrary orientation 67% of the time. We achieve a topline accuracy of 79% when aggregating predictions from ten patterns of the same material but different zone axes. The space group and lattice predictions range from 70% to 90% accuracy and median errors of 0.01-0.5Å, respectively, for cubic, hexagonal, trigonal, and tetragonal crystal systems while being less reliable on orthorhombic and monoclinic systems. We apply this architecture to a four-dimensional scanning transmission electron microscopy scan of gold nanoparticles, where it accurately predicts the crystal structure and lattice constants. These random forest models can be used to significantly accelerate the analysis of electron diffraction patterns, particularly in the case of unknown crystal structures. Additionally, due to the speed of inference, these models could be integrated into live transmission electron microscopy experiments, allowing real-Time labeling of a specimen.

36 MATERIALS SCIENCE↗
DOE OSTI2023

Fast Particle-based Anomaly Detection Algorithm with Variational Autoencoder

Model-agnostic anomaly detection is one of the promising approaches in the search for new beyond the standard model physics. In this paper, we present Set-VAE, a particle-based variational autoencoder (VAE) anomaly detection algorithm. We demonstrate a 2x signal efficiency gain compared with traditional subjettiness-based jet selection. Furthermore, with an eye to the future deployment to trigger systems, we propose the CLIP-VAE, which reduces the inference-time cost of anomaly detection by using the KL-divergence loss as the anomaly score, resulting in a 2x acceleration in latency and reducing the caching requirement.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2024

Portable Acceleration of CMS Computing Workflows with Coprocessors as a Service

Computing demands for large scientific experiments, such as the CMS experiment at the CERN LHC, will increase dramatically in the next decades. To complement the future performance increases of software running on central processing units (CPUs), explorations of coprocessor usage in data processing hold great potential and interest. Coprocessors are a class of computer processors that supplement CPUs, often improving the execution of certain functions due to architectural design choices. We explore the approach of Services for Optimized Network Inference on Coprocessors (SONIC) and study the deployment of this as-a-service approach in large-scale data processing. In the studies, we take a data processing workflow of the CMS experiment and run the main workflow on CPUs, while offloading several machine learning (ML) inference tasks onto either remote or local coprocessors, specifically graphics processing units (GPUs). With experiments performed at Google Cloud, the Purdue Tier-2 computing center, and combinations of the two, we demonstrate the acceleration of these ML algorithms individually on coprocessors and the corresponding throughput improvement for the entire workflow. This approach can be easily generalized to different types of coprocessors and deployed on local CPUs without decreasing the throughput performance. We emphasize that the SONIC approach enables high coprocessor usage and enables the portability to run workflows on different types of coprocessors.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2022

A machine learning approach to determine the elastic properties of printed fiber-reinforced polymers

This work focuses on the simultaneous determination of the elastic constants and the fiber orientation state for a short fiber-reinforced polymer composite by performing a minimum of experimental tests. Here we introduce a methodology that enables the inverse determination of fiber orientation state and the in-situ polymer properties by performing tensile tests at the composite coupon level. We demonstrate the approach for the extrusion deposition additive manufacturing (EDAM) process to illustrate one application of the methodology, but the development is such that it can be applied to short fiber-reinforced polymer (SFRP) systems processed via other methods. Currently, developing composites additive manufacturing digital twins require extensive material characterization. In particular, the mechanical characterization of the orthotropic elastic properties of a composite involves extensive sample preparation and testing, therefore the elasticity tensor is generally populated using a micromechanics model. This, however, requires measuring the fiber orientation state in addition to knowing the constituent material properties. Experimentally measuring the fiber orientation state can be tedious and time consuming. Further, optical methods are limited to resolving the orientation of cylindrical fibers or cluster of non-cylindrical fibers, and computed tomography (CT) methods scan regions of volume that are much smaller than a full printed bead. Therefore, we propose a methodology, accelerated by machine learning, to identify the anisotropic mechanical properties and fiber orientation state at the same time. Early results show that inference of the fiber orientation and composite properties is possible with as few as three tensile tests. Our results show that a combination of the choice of the micromechanics model and reliable set of experiments can yield the nine elastic constants, as well as, the fiber orientation state.

36 MATERIALS SCIENCE↗
DOE OSTI2021

Physics-informed neural networks for solving forward and inverse flow problems via the Boltzmann-BGK formulation

We report the Boltzmann equation with the Bhatnagar-Gross-Krook collision model (Boltzmann-BGK equation) has been employed to describe multiscale flows, i.e., from the hydrodynamic limit to free molecular flow. In this study we employ physics-informed neural networks (PINNs) to solve forward and inverse problems via the Boltzmann-BGK formulation (PINN-BGK), enabling PINNs to model flows in both the continuum and rarefied regimes. In particular, the PINN-BGK is composed of three sub-networks, i.e., the first for approximating the equilibrium distribution function, the second for approximating the non-equilibrium distribution function, and the third one for encoding the Boltzmann-BGK equation as well as the corresponding boundary/initial conditions. By minimizing the residuals of the governing equations and the mismatch between the predicted and provided boundary/initial conditions, we can approximate the Boltzmann-BGK equation for both continuous and rarefied flows. For forward problems, the PINN-BGK is utilized to solve various benchmark flows given boundary/initial conditions, e.g. Kovasznay flow, Taylor-Green flow, cavity flow, and micro Couette flow for Knudsen number up to 5. For inverse problems, we focus on rarefied flows in which accurate boundary conditions are difficult to obtain. We employ the PINN-BGK to infer the flow field in the entire computational domain given a limited number of interior scattered measurements on the velocity without using the (unknown) boundary conditions. Results for the two-dimensional micro Couette and micro cavity flows with Knudsen numbers ranging from 0.1 to 10 indicate that the PINN-BGK can infer the velocity field in the entire domain with good accuracy. Finally, we also present some results on using transfer learning to accelerate the training process. Specifically, we can obtain a three-fold speedup comparted to the standard training process (e.g., Adam plus L-BFG-B) for the two-dimensional flow problems considered in our work.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2026

AI-driven Neutrino Beam Diagnostics for Next-Generation Neutrino Experiments

The accelerator-driven beam uncertainty limits oscillation measurements in long-baseline neutrino experiments. Spill-resolved beam diagnostics and real-time inference are necessary to address these neutrino flux systematics. As such, we present a machine-learning-based beam monitoring framework developed and validated using data from the T2K experiment. Our approach uses downstream, spill-by-spill muon monitor observables to predict upstream parameters such as proton beam position and width. We achieve high predictive accuracy on nominal runs, demonstrating robust baseline performance whether the model is trained on stable runs or systematically varied conditions. The framework is designed to be robust against domain shifts, allowing the neural network architectures and inference strategies developed with T2K data to be retrained and validated using LBNF simulations, with the goal of eventual deployment under real LBNF/DUNE operating conditions. This scalable approach to real-time beam inference offers a pathway toward reducing flux systematics for next-generation neutrino experiments such as DUNE.

Aney, Noah [Fermilab; U. Chicago (main)]↗
DOE OSTI2026

Supercharging simulation-based inference for Bayesian optimal experimental design

Abstract Bayesian optimal experimental design (BOED) seeks to maximize the expected information gain (EIG) of experiments. This requires a likelihood estimate, which in many settings is intractable. Simulation-based inference (SBI) provides powerful tools for this regime. However, existing work explicitly connecting SBI and BOED is restricted to a single contrastive EIG bound. We show that the EIG admits multiple formulations which can directly leverage modern SBI density estimators, encompassing neural posterior, likelihood, and ratio estimation. Building on this perspective, we define a novel EIG estimator using neural likelihood estimation. Further, we identify optimization as a key bottleneck of gradient based EIG maximization and show that a simple multi-start parallel gradient ascent procedure can substantially improve reliability and performance. With these innovations, our SBI-based BOED methods are able to match or outperform by up to 22% existing state-of-the-art approaches across standard BOED benchmarks.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Electron Energization by Inertial Alfvén Waves in Density Depleted Flux Tubes at Jupiter

Juno satellite observations have illustrated substantial examples of broadband electron energization up to 10 5 –10 6 eV levels. In order to explain these observations, we use a hybrid gyrofluid kinetic-electron model in an untilted dipolar topology to illustrate energization to high levels in weak current conditions by inertial Alfvén waves close to the Jupiter ionosphere for ambient plasma densities and magnetic field perturbations inferred from Juno satellite observations. The key to the high energization is the extremely low densities evident in the observations which necessitates the acceleration of electrons to very high velocities in order to carry the field-aligned current.

58 GEOSCIENCES↗
DOE OSTI2024

Improvement and generalization of ABCD method with Bayesian inference

To find New Physics or to refine our knowledge of the Standard Model at the LHC is an enterprise that involves many factors, such as the capabilities and the performance of the accelerator and detectors, the use and exploitation of the available information, the design of search strategies and observables, as well as the proposal of new models. We focus on the use of the information and pour our effort in re-thinking the usual data-driven ABCD method to improve it and to generalize it using Bayesian Machine Learning techniques and tools. We propose that a dataset consisting of a signal and many backgrounds is well described through a mixture model. Signal, backgrounds and their relative fractions in the sample can be well extracted by exploiting the prior knowledge and the dependence between the different observables at the event-by-event level with Bayesian tools. We show how, in contrast to the ABCD method, one can take advantage of understanding some properties of the different backgrounds and of having more than two independent observables to measure in each event. In addition, instead of regions defined through hard cuts, the Bayesian framework uses the information of continuous distribution to obtain soft-assignments of the events which are statistically more robust. To compare both methods we use a toy problem inspired by pp\to hh\to b\bar b b \bar b p p → h h → b b ‾ b b ‾ , selecting a reduced and simplified number of processes and analysing the flavor of the four jets and the invariant mass of the jet-pairs, modeled with simplified distributions. Taking advantage of all this information, and starting from a combination of biased and agnostic priors, leads us to a very good posterior once we use the Bayesian framework to exploit the data and the mutual information of the observables at the event-by-event level. We show how, in this simplified model, the Bayesian framework outperforms the ABCD method sensitivity in obtaining the signal fraction in scenarios with 1% and 0.5% true signal fractions in the dataset. We also show that the method is robust against the absence of signal. We discuss potential prospects for taking this Bayesian data-driven paradigm into more realistic scenarios.

Alvarez, Ezequiel↗
DOE OSTI2025

Unveiling X-ray absorption signatures of boron nitride via first-principles simulation and machine learning

Boron nitride (BN) allotropes hold great promise in many advanced applications ranging from optical and photonic devices to energy storage and battery systems to tribological components. The diverse functionalities of this material stem from BN’s highly tunable structural and electronic properties, which are governed by the versatile boron–nitrogen bonding configurations. Exploring the structural landscape of BN can unveil novel structures possessing unique properties suited for specific applications, therefore accelerating the design of next-generation advanced functional materials. In this work, we leverage boron K-edge X-ray absorption spectroscopy (XAS) as an effective probe for local structural features and chemical environments. A total of 210 BN crystal structures are generated via analogies to the extensive array of carbon allotropes, and XAS is simulated for each unique local motif within the resulting collection of structures. A mapping between structural features and spectral signatures was established by synergizing first-principle simulations with data-driven based post-analysis approaches. Specifically, we developed a neural network model that can satisfactorily predict spectra line shapes from local structural descriptors. Toward automatic spectroscopic interpretation of any new BN structures, supervised machine learning models, trained on this structure–spectrum dataset, can accurately infer local coordination environments from simulated XAS, highlighting the strength of this unique approach of combining high-fidelity first-principles simulation and machine-learning to accelerate target design of novel BN materials via rational understanding of local structure-spectrum correlations.

36 MATERIALS SCIENCE↗
DOE OSTI2023

Advanced data analysis in inertial confinement fusion and high energy density physics

Bayesian analysis enables flexible and rigorous definition of statistical model assumptions with well-characterized propagation of uncertainties and resulting inferences for single-shot, repeated, or even cross-platform data. This approach has a strong history of application to a variety of problems in physical sciences ranging from inference of particle mass from multi-source high-energy particle data to analysis of black-hole characteristics from gravitational wave observations. The recent adoption of Bayesian statistics for analysis and design of high-energy density physics (HEDP) and inertial confinement fusion (ICF) experiments has provided invaluable gains in expert understanding and experiment performance. In this Review, we discuss the basic theory and practical application of the Bayesian statistics framework. We highlight a variety of studies from the HEDP and ICF literature, demonstrating the power of this technique. Due to the computational complexity of multi-physics models needed to analyze HEDP and ICF experiments, Bayesian inference is often not computationally tractable. Two sections are devoted to a review of statistical approximations, efficient inference algorithms, and data-driven methods, such as deep-learning and dimensionality reduction, which play a significant role in enabling use of the Bayesian framework. We provide additional discussion of various applications of Bayesian and machine learning methods that appear to be sparse in the HEDP and ICF literature constituting possible next steps for the community. We conclude by highlighting community needs, the resolution of which will improve trust in data-driven methods that have proven critical for accelerating the design and discovery cycle in many application areas.

47 OTHER INSTRUMENTATION↗
DOE OSTI2023

Detecting Permafrost Active Layer Thickness Change From Nonlinear Baseflow Recession

We report permafrost underlies about one fifth of the global land area and affects ground stability, freshwater runoff, soil chemistry, and surface-atmosphere gas exchange. The depth of thawed ground overlying permafrost (active layer thickness) has broadly increased across the Arctic in recent decades, coincident with a period of increased streamflow, especially the lowest flows (baseflow). Mechanistic links between active layer thickness and baseflow have recently been explored using linear reservoir theory, but most watersheds behave as nonlinear reservoirs. We derive theoretical nonlinear relationships between long-term average saturated soil thickness $\bar{η}$(proxy for active layer thickness) and long-term average baseflow. When applied to 38 years of daily streamflow data for the Kuparuk River basin on the North Slope of Alaska, the theory predicts $\bar{η}$ increased 0.17 ± 0.22[2σ] cm a -1 between 1983 and 2020 (6.4 ± cm total). The rate of increase nearly doubled to 0.29 ± 0.31 cm a -1 between 1990 and 2020, during which time local field measurements from Circumpolar Active Layer Monitoring sites indicate the active layer increased 0.31 ± 0.22 cm a -1 . The predicted rate of increase more than doubled again between 2002 and 2020, outpacing a near doubling of observed active layer thickening, consistent with trends in terrestrial water storage inferred from Gravity Recovery and Climate Experiment satellite gravimetry and Modern-Era Retrospective Analysis for Research and Applications climate reanalysis. Overall, hydrologic change is accelerating in the Kuparuk River basin, and we provide a theoretical framework for estimating basin-scale changes in active layer water storage from streamflow measurements.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Platform Of Optimal Experiment Management

The platform of optimal experiment management, POEM, powered with automated machine learning to accelerate the discovery of optimal solutions, and automatically guide the design of experiments to be evaluated. POEM currently supports 1) random model explorations for experiment design, 2) sparse grid model explorations with Gaussian Polynomial Chaos surrogate model to accelerate experiment design ,3) time-dependent model sensitivity and uncertainty analysis to identify the importance features for experiment design, 4) model calibrations via Bayesian inference to integrate experiments to improve model performance, and 5) Bayesian optimization for optimal experimental design. In addition, POEM aims to simplify the process of experimental design for users, enabling them to analyze the data with minimal human intervention, and improving the technological output from research activities.

Wang, Congjian [Idaho National Laboratory (INL), I↗
DOE OSTI2021

Accelerating Transformer-based Deep Learning Models on FPGAs using Column Balanced Block Pruning

Although Transformer-based language representations achieve state-of-the-art accuracy on various natural language processing (NLP) tasks, the large model size has been challenging the resource constrained computing platforms. Weight pruning, as a popular and effective technique in reducing the number of weight parameters and accelerating the Transformer, has been investigated on GPUs. However, the Transformer acceleration using weight pruning on field-programmable gate array (FPGAs) remains unexplored. This paper investigates the column balanced block-wise pruning on Transformer and designs an FPGA acceleration engine to customize the balanced blockwise matrix multiplication. We implement the Transformer model with proper hardware scheduling, and the experiments show that the Transformer inference on FPGA achieves 10.35 ms latency with the batch size of 32, which is 10.96 × speed up comparing to CPU platform and 2.08 × speed up comparing to GPU platform.

Peng, Hongwu↗
DOE OSTI2021

Characterisation of a laser plasma betatron source for high resolution x-ray imaging

Abstract We report on the characterisation of an x-ray source, generated by a laser-driven plasma wakefield accelerator. The spectrum of the optimised source was consistent with an on-axis synchrotron spectrum with a critical energy of 13.8 − 1.9 + 2.2 keV and the number of photons per pulse generated above 1 keV was calculated to be 6 − 0.9 + 1.2 × 10 9 . The x-ray beam was used to image a resolution grid placed 37 cm from the source, which gave a measured spatial resolution of 4 µ m × 5 µ m. The inferred emission region had a radius and length of 0.5 ± 0.2 µ m and 3.2 ± 0.9 mm respectively. It was also observed that laser damage to the exit aperture of the gas cell led to a reduction in the accelerated electron beam charge and a corresponding reduction in x-ray flux due to the change in the plasma density profile.

plasma acceleration↗
DOE OSTI2025

Lens Modeling of STRIDES Strongly Lensed Quasars Using Neural Posterior Estimation

Strongly lensed quasars can be used to constrain cosmological parameters through time-delay cosmography. Models of the lens masses are a necessary component of this analysis. To enable time-delay cosmography from a sample of $\mathcal{O}(10^3)$ lenses, which will soon become available from surveys like the Rubin Observatory’s Legacy Survey of Space and Time and the Euclid Wide Survey, we require fast and standardizable modeling techniques. To address this need, we apply neural posterior estimation (NPE) for modeling galaxy-scale strongly lensed quasars from the Strong Lensing Insights into the Dark Energy Survey (STRIDES) sample. NPE brings two advantages: speed and the ability to implicitly marginalize over nuisance parameters. We extend this method by employing sequential NPE to increase precision of mass model posteriors. We then fold individual lens models into a hierarchical Bayesian inference to recover the population distribution of lens mass parameters, accounting for out-of-distribution shift. After verifying our method using simulated analogs of the STRIDES lens sample, we apply our method to 14 Hubble Space Telescope single-filter observations. We find the population mean of the power-law elliptical mass distribution slope, γ lens , to be $\mathcal{M}_γ$ lens = 2.13 ± 0.06. Our result represents the first population-level constraint for these systems. This population-level inference from fully automated modeling is an important stepping stone toward cosmological inference with large samples of strongly lensed quasars.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2025

Pathway to tune sputter coated Nb$_3$Sn with Zr

Superconducting radio frequency (SRF) cavities are essential elements of many modern-day particle accelerators such as, European XFEL (DESY), LCLS II-HE (SLAC) and PIP II (Fermilab). Nb$_3$Sn superconducting material promises significant potential to exceed the performance of conventional niobium based SRF cavities. We aim to advance the ongoing R$&$D efforts to achieve Nb$_3$Sn coated cavities with high material quality. In this work, we have developed to achieve a route for effectively reducing the intrinsic defects in magnetron sputter coated Nb$_3$Sn SRF cavities and eventually improve the overall RF performance. We have observed that Zr is mainly being incorporated as ZrO$_2$ precipitates of average dimensions ranging from 20-100nm. By increasing the concentration of Zr, we have three major effects in the physical properties. Firstly, we noted that the density of the surface and bulk voids as well as their average dimensions are dramatically reduced on increasing the Zr content in Nb$_3$Sn. These surface voids are believed to be detrimental to cavity applications due to increase the local dissipation of electric and magnetic fields. These results indicate that inclusion of Zr in Nb$_3$Sn sputtered coating will improve the material quality and might help to reduce the overall RF power dissipation. Secondly, we observed that increasing Zr concentration from 0 to 2.4% substantially improves both superconducting transition temperature (Tc) and upper critical magnetic field (Hc2). However, on further increasing the Zr concentration to 24%, we noticed a small decrement in the Hc2 value. These results infer that the increase in Zr concentration up to an optimal concentration will be beneficial to enhance the superconducting properties. Increase in Tc promises lesser RF losses and improvement in accelerating gradient, whereas the increase in Hc2 suggests more efficient flux pinning which could give rise to high quality factors in presence of multi-Tesla magnetic field. Additionally, we also observed that increasing Zr concentration possibly prevents the oxygen diffusion and leads to thinner formation of primary surface oxide (Nb$_2$O$_5$ and SnO). This study gives a route to tune Zr fraction in Nb3Sn to modify the physical properties that control the high field performance of Nb3Sn RF cavities.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗