Search NASA⌕ Search

SEARCH · Search NASA

Results for “gradient methods”

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

Aerodynamic optimization of supersonic compressor cascade using differential evolution on GPU

Differential Evolution (DE) is a powerful stochastic optimization method. Compared to gradient-based algorithms, DE is able to avoid local minima but requires at the same time more function evaluations. In turbomachinery applications, function evaluations are performed with time-consuming CFD simulation, which results in a long, non affordable, design cycle. Modern High Performance Computing systems, especially Graphic Processing Units (GPUs), are able to alleviate this inconvenience by accelerating the design evaluation itself. In this work we present a validated CFD Solver running on GPUs, able to accelerate the design evaluation and thus the entire design process. An achieved speedup of 20x to 30x enabled the DE algorithm to run on a high-end computer instead of a costly large cluster. The GPU-enhanced DE was used to optimize the aerodynamics of a supersonic compressor cascade, achieving an aerodynamic loss minimization of 20%.

97 MATHEMATICS AND COMPUTING↗

THE SPECTROSCOPIC DIVERSITY OF TYPE Ia SUPERNOVAE

We present 2603 spectra of 462 nearby Type Ia supernovae (SNe Ia), including 2065 previously unpublished spectra, obtained during 1993-2008 through the Center for Astrophysics Supernova Program. There are on average eight spectra for each of the 313 SNe Ia with at least two spectra. Most of the spectra were obtained with the FAST spectrograph at the Fred Lawrence Whipple Observatory 1.5 m telescope and reduced in a consistent manner, making this data set well suited for studies of SN Ia spectroscopic diversity. Using additional data from the literature, we study the spectroscopic and photometric properties of SNe Ia as a function of spectroscopic class using the classification schemes of Branch et al. and Wang et al. The width-luminosity relation appears to be steeper for SNe Ia with broader lines, although the result is not statistically significant with the present sample. Based on the evolution of the characteristic Si II {lambda}6355 line, we propose improved methods for measuring velocity gradients, revealing a larger range than previously suspected, from {approx}0 to {approx}400 km s{sup -1} day{sup -1} considering the instantaneous velocity decline rate at maximum light. We find a weaker and less significant correlation between Si II velocity and intrinsic B - V color at maximum light than reported by Foley et al., owing to a more comprehensive treatment of uncertainties and host galaxy dust. We study the extent of nuclear burning and the presence of unburnt carbon in the outermost layers of the ejecta and report new detections of C II {lambda}6580 in 23 early-time SN Ia spectra. The frequency of C II detections is not higher in SNe Ia with bluer colors or narrower light curves, in conflict with the recent results of Thomas et al. Based on nebular spectra of 27 SNe Ia, we find no relation between the FWHM of the iron emission feature at {approx}4700 A and {Delta}m{sub 15}(B) after removing the two low-luminosity SN 1986G and SN 1991bg, suggesting that the peak luminosity is not strongly dependent on the kinetic energy of the explosion for most SNe Ia. Finally, we confirm the correlation of velocity shifts in some nebular lines with the intrinsic B - V color of SNe Ia at maximum light, although several outliers suggest a possible non-monotonic behavior for the largest blueshifts.

79 ASTRONOMY AND ASTROPHYSICS↗

Degenerate doping in B-Ga 2 O 3 Single Crystals through Hf-doping

n-type conductivity of β-Ga 2 O 3 grown from the melt is typically achieved using Sn and Si. In this paper, we experimentally and computationally investigate Hf doping of β-Ga 2 O 3 single crystals using UV–vis-NIR absorption and Hall effect measurements and hybrid functional calculations. Unintentionally-doped and Hf-doped samples with a nominal concentration of 0.5at% were grown from the melt using vertical gradient freeze and Czochralski method in mixed Ar + O 2 atmosphere. We demonstrate Hf dopants, predicted to incorporate on the octahedral GaII site as a shallow donor, achieve degenerate doping in β-Ga 2 O 3 with a measured electron concentration ~2 × 10 19 cm -3 , mobility 80–65 cm 2 V -1 s -1 , and resistivity down to 5 mΩ cm in our samples. The concentration of Hf was measured to be 1.3 × 10 19 atoms cm -3 using glow discharge mass spectroscopy on doped samples, confirming Hf to be the cause of n-type conductivity (electron concentration ~2 × 10 19 cm -3 ).

36 MATERIALS SCIENCE↗

A log-barrier Newton-CG method for bound constrained optimization with complexity guarantees

Abstract We describe an algorithm based on a logarithmic barrier function, Newton’s method and linear conjugate gradients that seeks an approximate minimizer of a smooth function over the non-negative orthant. We develop a bound on the complexity of the approach, stated in terms of the required accuracy and the cost of a single gradient evaluation of the objective function and/or a matrix-vector multiplication involving the Hessian of the objective. The approach can be implemented without explicit calculation or storage of the Hessian.

O’Neill, Michael↗

Infrared fixed point of the SU(3) gauge theory with 𝑁 𝑓 = 10 flavors

We use lattice simulations and the continuous renormalization group method, based on the gradient flow, to calculate the 𝛽 function and anomalous dimensions of the SU(3) gauge theory with 𝑁 𝑓 = 10 flavors of fermions in the fundamental representation. We employ several improvements to extend the range of available renormalized couplings, including the addition of heavy Pauli-Villars bosons to reduce cutoff effects and the combination of a range of gradient flow transformations. While in the weak coupling regime our result is consistent with those of earlier studies, our techniques allow us to study the system at much stronger couplings than previously possible. We find that the renormalization group 𝛽 function develops a zero, corresponding to an infrared-stable fixed point, at gradient-flow coupling 𝑔 2 = 15.0⁢(5). We also determine the mass and tensor anomalous dimensions: At the fixed point we find 𝛾 𝑚 ≃0.6, suggesting that this system might be deep inside the conformal window.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

AlignOT: An Optimal Transport Based Algorithm for Fast 3D Alignment With Applications to Cryogenic Electron Microscopy Density Maps

Aligning electron density maps from Cryogenic electron microscopy (cryo-EM) is a first key step for studying multiple conformations of a biomolecule. As this step remains costly and challenging, with standard alignment tools being potentially stuck in local minima, we propose here a new procedure, called AlignOT, which relies on the use of computational optimal transport (OT) to align EM maps in 3D space. By embedding a fast estimation of OT maps within a stochastic gradient descent algorithm, our method searches for a rotation that minimizes the Wasserstein distance between two maps, represented as point clouds. Here, we quantify the impact of various parameters on the precision and accuracy of the alignment, and show that AlignOT can outperform the standard local alignment methods, with an increased range of rotation angles leading to proper alignment. We further benchmark AlignOT on various pairs of experimental maps, which account for different types of conformational heterogeneities and geometric properties. As our experiments show good performance, we anticipate that our method can be broadly applied to align 3D EM maps.

3D alignment↗

Theoretical study of oxygen sorption and diffusion in the volume and on the surface of a γ-TiAl alloy

The oxygen sorption on the low-index (001), (100), and (110) surfaces of a γ-TiAl alloy is studied by the pseudopotential method with the generalized gradient approximation for the exchange-correlation functional. The most preferred sites for oxygen sorption in the bulk and on the surface of the alloy are determined. The titanium-rich octahedral site is shown to be preferred for oxygen sorption in the bulk material. The effect of the oxygen concentration on the atomic and electronic structures of the stoichiometric TiAl(100) surface is studied. It is shown that, at the first stage of oxidation, oxygen prefers to form bonds with titanium. The energy barriers for oxygen diffusion on the stoichiometric (100) surface and in the bulk of the material are calculated. The energy barriers are shown to depend substantially on the local environments of oxygen and to increase during diffusion from titanium-rich sites. The most possible mechanism of oxygen diffusion from the (100) surface to the bulk of the material is oxygen migration through tetrahedral sites.

36 MATERIALS SCIENCE↗

A Framework for Error-Bounded Approximate Computing, with an Application to Dot Products

Approximate computing techniques, which trade off the computation accuracy of an algorithm for better performance and energy efficiency, have been successful in reducing computation and power costs in several domains. However, error sensitive applications in high-performance computing are unable to benefit from existing approximate computing strategies that are not developed with guaranteed error bounds. While approximate computing techniques can be developed for individual high-performance computing applications by domain specialists, this often requires additional theoretical analysis and potentially extensive software modification. Hence, the development of low-level error-bounded approximate computing strategies that can be introduced into any high-performance computing application without requiring additional analysis or significant software alterations is desirable. In this paper, we provide a contribution in this direction by proposing a general framework for designing error-bounded approximate computing strategies and apply it to the dot product kernel to develop \bf qdot---an error-bounded approximate dot product kernel. Following the introduction of qdot, here we perform a theoretical analysis that yields a deterministic bound on the relative approximation error introduced by qdot. Empirical tests are performed to illustrate the tightness of the derived error bound and to demonstrate the effectiveness of qdot on a synthetic dataset, as well as two scientific benchmarks---the conjugate gradient (CG) and power methods. In some instances, using qdot for the dot products in CG can result in many components being quantized to half precision without increasing the iteration count required for convergence to the same solution as CG using a double precision dot product.

97 MATHEMATICS AND COMPUTING↗

Gradient Flow Renormalization of Quasi-Parton Distributions

Gradient flow, is a method that allows us to construct lattice quasi PDFs that are finite in the continuum limit. In the small flow time limit, the moments of the smeared quasi PDFs are proportional to those of the light-front PDFs. These proportionality coefficient obey a renormalization group equation that allows us to derive renormalization scale evolution equations for the matching kernel that relates the smeared quasi PDF and the renormalized light-front PDFs. Based on these results we propose a new methodology for obtaining light-front PDFs from lattice QCD.

Orginos, K.↗

Search for HH → bbτ⁺τ⁻ Using Run 3 Scouting Data Analyze b-tagging and tau-tagging Performance with Unified Particle Transformer

B-tagging and tau-tagging performances play an important role in the search for the rare event HH → bbτ⁺τ⁻. A transformer-based neural network, Unified Particle Transformer, is applied for both tagging tasks, and Run 3 proton–proton collision scouting data at center-of-mass energy of 13.6 TeV is used. The scouting data stream accepts events at a much higher rate compared to traditional triggers, but stores only the objects reconstructed in the trigger, no low-level detector information. Therefore, existing taggers trained for the offline event reconstruction cannot be used. Analysis of the SoftMax plots, ROC/AUC curves, confusion matrix, accuracy and losses are used to evaluate model performance. Specifically, the tagging efficiency of the signal and misidentification probability across multiple background processes are compared for varying working points. Different training samples with distinct distributions of jet flavors are utilized and related model performances are analyzed. Interpretability methods, such as Integrated Gradients, may further be applied to study the input features’ influence on the model’s decisions, providing insights into potential improvements.

Chen, Blair [Purdue U., West Lafayette; Fermilab]↗

Numerical simulations of liquid jetting with solid inclusions

The dynamics of finite-sized particles in fluids, and their influence on the overall flow, are of great interest across several industrial, environmental, and medical fields. In the context of inkjet printing, the presence of solid inclusions can be either intentional, as in additive manufacturing, or unintentional, as in standard printing processes. These inclusions can strongly impact the jetting process, causing effects such as jet asymmetry, bubble entrapment, and the formation of satellite droplets. Understanding and controlling particle behavior is therefore essential, particularly to predict how and when particles are ejected over multiple jetting cycles. It is therefore critical to develop reliable models that allow for a deeper understanding of the complex interplay between particle and fluid during the whole printing process. To address this, we present a tailored implementation of the Color-Gradient multicomponent Lattice Boltzmann Method for fully resolved three-dimensional (3D) simulations of multicycle liquid jetting with particles. Our method supports realistic parameter settings aligned with industrial inkjet systems, and we provide both qualitative and quantitative validation against experimental data. Additionally, we introduce a simplified model based on the Stokes drag law, in which solid particles are represented as point particles and do not influence the fluid flow. Despite this limitation, the model offers a computationally efficient means to explore the vast parameter space typically encountered in industrial applications, allowing, e.g., identifying critical ejection regions and estimating the number of cycles required for particle release. These qualitative insights are valuable for guiding and complement fully two-way coupled simulations.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

HGQ: High Granularity Quantization for Real-time Neural Networks on FPGAs

Neural networks with sub-microsecond inference latency are required by many critical applications. Targeting such applications deployed on FPGAs, we present High Granularity Quantization (HGQ), a quantization-aware training framework that optimizes parameter bit-widths through gradient descent. Unlike conventional methods, HGQ determines the optimal bit-width for each parameter independently, making it suitable for hardware platforms supporting heterogeneous arbitrary precision arithmetic. In our experiments, HGQ shows superior performance compared to existing network compression methods, achieving orders of magnitude reduction in resource consumption and latency while maintaining the accuracy on several benchmark tasks. These improvements enable the deployment of complex models previously infeasible due to resource or latency constraints. HGQ is open-source and is used for developing next-generation trigger systems at the CERN ATLAS and CMS experiments for particle physics, enabling the use of advanced machine learning models for real-time data selection with sub-microsecond latency.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Apparatus and process for passivating an SRF cavity

An apparatus and process for the production of a niobium cavity exhibiting high quality factors at high gradients is provided. The apparatus comprises a first chamber positioned within a second chamber, an RF generator and vacuum pumping systems. The process comprises placing the niobium cavity in a first chamber of the apparatus; thermally treating the cavity by high temperature in the first chamber while maintaining high vacuum in the first and second chambers; and applying a passivating thin film layer to a surface of the cavity in the presence of a gaseous mixture and an RF field. Further a niobium cavity exhibiting high quality factors at high gradients produced by the method of the invention is provided.

36 MATERIALS SCIENCE↗

A Bayesian Learning Approach to Wireless Outdoor Heatmap Construction using Deep Gaussian Process

We present a novel Bayesian learning approach to outdoor radio heatmap construction utilizing deep Gaussian process (GP). The proposed approach employs a two-layer hierarchy which consists of two cascaded Gaussian processes that are capable of modeling more complex input-output relations than standard single-layer Gaussian processes. Since deriving the exact model likelihood is challenging, a lower bound is optimized instead so that gradient descent-based methods can be performed to find out the optimal model parameters. Typically, inducing points are used in GPs to facilitate low-rank approximation of covariance (kernel) matrices for computation speedup. However, the inaccuracy induced by inducing points can accumulate when stacking multiple layers of GP which may hinder the performance of deep GP. Moreover, since inducing points need to be learned, having them at all layers of deep GP also incurs computational burden. To overcome the above challenges, in contrast to the canonical deep GP model, we use a modified architecture where a full standard GP resides in the first layer and inducing points are only introduced for the second layer. This modified architecture strikes a balance between model accuracy and training complexity. In the proposed model, the noise parameter of the first GP layer is also eliminated to improve the training efficiency as the noise parameter at the output of the second layer suffices to model the uncertainty in the output. The proposed approach is evaluated on real-world datasets, in the form of location-Received Signal Strength (RSS) pairs, collected from the Platform for Open Wireless Data-driven Experimental Research (POWDER) located at the campus of the University of Utah. Experiment results show that the proposed approach can achieve smaller prediction errors on various training and testing data configurations than DNN-based and GP-based methods.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

The effect of carbon contents on intragranular ferrite formed in the V-Ti-N microalloyed steel with a carbon content gradient prepared by controlling the surface decarburization

Highlights: • A method for quick preparation of steels with different carbon contents was provided. • A steel containing carbon in gradient was obtained through surface decarburization. • The effect of carbon contents on IGF nucleation was studied with the new method. - Abstract: The V-Ti-N microalloyed steel with a carbon content gradient was innovatively obtained through surface decarburization. Then, the effect of carbon contents on intragranular ferrite (IGF) formation was studied. With the carbon contents increasing from 0.39% to 0.47% (mass percent), the volume fraction of IGF decreases from 8.2% to 4.6%, but the number of IGF has no significant change. The critical size of IGF nuclei was calculated on the basis of classic heterogeneous nucleation theory. According to the results of calculation and statistics, with the carbon content increasing from 0.39% to 0.47%, the number of IGF nuclei does not decrease significantly. When the carbon content is relatively low, some (Ti, V)(C, N) particles precipitating in austenite matrix may also act as the nuclei of IGF to refine the pearlite-ferrite microstructure.

36 MATERIALS SCIENCE↗

An extrapolation method to remove spurious stress concentration in micromechanical analyses of composites using pixel-based meshes

It is easier to generate a finite element mesh of complicated geometries with voxel-based meshes compared to conformal meshes. However, the local stresses in voxel-based meshes are higher owing to artificial stress concentration caused due to jagged lines that approximate a curve. Here, we propose a method that accounts for this anomaly which would lead to accurate prediction of maximum von Mises stress in voxel-based meshes. The method is based on the variation of reciprocal of stresses and associated numerical gradient. The prediction of the proposed method is compared against the results of a conformal mesh by considering four two-dimensional example problems. The accuracy of the proposed method is verified using different mesh densities, material properties, and integration schemes. The proposed method was also implemented on a conformal mesh to verify that it filters out just the spurious stress concentration and not the realistic stress concentration.

36 MATERIALS SCIENCE↗